Downloads: 1
Diff of
/datasets/cdsl/preprocess.ipynb
[000000]
..
[d6904d]
Switch to unified view
| a | b/datasets/cdsl/preprocess.ipynb | ||
|---|---|---|---|
| 1 | { |
||
| 2 | "cells": [ |
||
| 3 | { |
||
| 4 | "cell_type": "markdown", |
||
| 5 | "metadata": {}, |
||
| 6 | "source": [ |
||
| 7 | "# hm dataset pre-processing\n", |
||
| 8 | "\n", |
||
| 9 | "import packages" |
||
| 10 | ] |
||
| 11 | }, |
||
| 12 | { |
||
| 13 | "cell_type": "code", |
||
| 14 | "execution_count": null, |
||
| 15 | "metadata": {}, |
||
| 16 | "outputs": [], |
||
| 17 | "source": [ |
||
| 18 | "import os\n", |
||
| 19 | "import pandas as pd\n", |
||
| 20 | "import numpy as np\n", |
||
| 21 | "import matplotlib.pyplot as plt\n", |
||
| 22 | "import pickle as pkl\n", |
||
| 23 | "import torch\n", |
||
| 24 | "import math\n", |
||
| 25 | "import datetime\n", |
||
| 26 | "from tqdm import tqdm\n", |
||
| 27 | "import datetime\n", |
||
| 28 | "import re\n", |
||
| 29 | "from functools import reduce" |
||
| 30 | ] |
||
| 31 | }, |
||
| 32 | { |
||
| 33 | "cell_type": "markdown", |
||
| 34 | "metadata": {}, |
||
| 35 | "source": [ |
||
| 36 | "## Demographic data" |
||
| 37 | ] |
||
| 38 | }, |
||
| 39 | { |
||
| 40 | "cell_type": "code", |
||
| 41 | "execution_count": null, |
||
| 42 | "metadata": {}, |
||
| 43 | "outputs": [], |
||
| 44 | "source": [ |
||
| 45 | "demographic = pd.read_csv('./raw_data/19_04_2021/COVID_DSL_01.CSV', encoding='ISO-8859-1', sep='|')\n", |
||
| 46 | "print(len(demographic))\n", |
||
| 47 | "demographic.head()" |
||
| 48 | ] |
||
| 49 | }, |
||
| 50 | { |
||
| 51 | "cell_type": "code", |
||
| 52 | "execution_count": null, |
||
| 53 | "metadata": {}, |
||
| 54 | "outputs": [], |
||
| 55 | "source": [ |
||
| 56 | "med = pd.read_csv('./raw_data/19_04_2021/COVID_DSL_04.CSV', encoding='ISO-8859-1', sep='|')\n", |
||
| 57 | "print(len(med))\n", |
||
| 58 | "med.head()" |
||
| 59 | ] |
||
| 60 | }, |
||
| 61 | { |
||
| 62 | "cell_type": "code", |
||
| 63 | "execution_count": null, |
||
| 64 | "metadata": {}, |
||
| 65 | "outputs": [], |
||
| 66 | "source": [ |
||
| 67 | "len(med['ID_ATC7'].unique())" |
||
| 68 | ] |
||
| 69 | }, |
||
| 70 | { |
||
| 71 | "cell_type": "markdown", |
||
| 72 | "metadata": {}, |
||
| 73 | "source": [ |
||
| 74 | "get rid of patient with missing label" |
||
| 75 | ] |
||
| 76 | }, |
||
| 77 | { |
||
| 78 | "cell_type": "code", |
||
| 79 | "execution_count": null, |
||
| 80 | "metadata": {}, |
||
| 81 | "outputs": [], |
||
| 82 | "source": [ |
||
| 83 | "print(len(demographic))\n", |
||
| 84 | "demographic = demographic.dropna(axis=0, how='any', subset=['IDINGRESO', 'F_INGRESO_ING', 'F_ALTA_ING', 'MOTIVO_ALTA_ING'])\n", |
||
| 85 | "print(len(demographic))" |
||
| 86 | ] |
||
| 87 | }, |
||
| 88 | { |
||
| 89 | "cell_type": "code", |
||
| 90 | "execution_count": null, |
||
| 91 | "metadata": {}, |
||
| 92 | "outputs": [], |
||
| 93 | "source": [ |
||
| 94 | "def outcome2num(x):\n", |
||
| 95 | " if x == 'Fallecimiento':\n", |
||
| 96 | " return 1\n", |
||
| 97 | " else:\n", |
||
| 98 | " return 0\n", |
||
| 99 | "\n", |
||
| 100 | "def to_one_hot(x, feature):\n", |
||
| 101 | " if x == feature:\n", |
||
| 102 | " return 1\n", |
||
| 103 | " else:\n", |
||
| 104 | " return 0" |
||
| 105 | ] |
||
| 106 | }, |
||
| 107 | { |
||
| 108 | "cell_type": "code", |
||
| 109 | "execution_count": null, |
||
| 110 | "metadata": {}, |
||
| 111 | "outputs": [], |
||
| 112 | "source": [ |
||
| 113 | "# select necessary columns from demographic\n", |
||
| 114 | "demographic = demographic[\n", |
||
| 115 | " [\n", |
||
| 116 | " 'IDINGRESO', \n", |
||
| 117 | " 'EDAD',\n", |
||
| 118 | " 'SEX',\n", |
||
| 119 | " 'F_INGRESO_ING', \n", |
||
| 120 | " 'F_ALTA_ING', \n", |
||
| 121 | " 'MOTIVO_ALTA_ING', \n", |
||
| 122 | " 'ESPECIALIDAD_URGENCIA', \n", |
||
| 123 | " 'DIAG_URG'\n", |
||
| 124 | " ]\n", |
||
| 125 | " ]\n", |
||
| 126 | "\n", |
||
| 127 | "# rename column\n", |
||
| 128 | "demographic = demographic.rename(columns={\n", |
||
| 129 | " 'IDINGRESO': 'PATIENT_ID',\n", |
||
| 130 | " 'EDAD': 'AGE',\n", |
||
| 131 | " 'SEX': 'SEX',\n", |
||
| 132 | " 'F_INGRESO_ING': 'ADMISSION_DATE',\n", |
||
| 133 | " 'F_ALTA_ING': 'DEPARTURE_DATE',\n", |
||
| 134 | " 'MOTIVO_ALTA_ING': 'OUTCOME',\n", |
||
| 135 | " 'ESPECIALIDAD_URGENCIA': 'DEPARTMENT_OF_EMERGENCY',\n", |
||
| 136 | " 'DIAG_URG': 'DIAGNOSIS_AT_EMERGENCY_VISIT'\n", |
||
| 137 | "})\n", |
||
| 138 | "\n", |
||
| 139 | "# SEX: male: 1; female: 0\n", |
||
| 140 | "demographic['SEX'].replace('MALE', 1, inplace=True)\n", |
||
| 141 | "demographic['SEX'].replace('FEMALE', 0, inplace=True)\n", |
||
| 142 | "\n", |
||
| 143 | "# outcome: Fallecimiento(dead): 1; others: 0\n", |
||
| 144 | "demographic['OUTCOME'] = demographic['OUTCOME'].map(outcome2num)\n", |
||
| 145 | "\n", |
||
| 146 | "# diagnosis at emergency visit (loss rate < 10%)\n", |
||
| 147 | "# demographic['DIFFICULTY_BREATHING'] = demographic['DIAGNOSIS_AT_EMERGENCY_VISIT'].map(lambda x: to_one_hot(x, 'DIFICULTAD RESPIRATORIA')) # 1674\n", |
||
| 148 | "# demographic['SUSPECT_COVID'] = demographic['DIAGNOSIS_AT_EMERGENCY_VISIT'].map(lambda x: to_one_hot(x, 'SOSPECHA COVID-19')) # 960\n", |
||
| 149 | "# demographic['FEVER'] = demographic['DIAGNOSIS_AT_EMERGENCY_VISIT'].map(lambda x: to_one_hot(x, 'FIEBRE')) # 455\n", |
||
| 150 | "\n", |
||
| 151 | "# department of emergency (loss rate < 10%)\n", |
||
| 152 | "# demographic['EMERGENCY'] = demographic['DEPARTMENT_OF_EMERGENCY'].map(lambda x: to_one_hot(x, 'Medicina de Urgencias')) # 3914" |
||
| 153 | ] |
||
| 154 | }, |
||
| 155 | { |
||
| 156 | "cell_type": "code", |
||
| 157 | "execution_count": null, |
||
| 158 | "metadata": {}, |
||
| 159 | "outputs": [], |
||
| 160 | "source": [ |
||
| 161 | "# del useless data\n", |
||
| 162 | "demographic = demographic[\n", |
||
| 163 | " [\n", |
||
| 164 | " 'PATIENT_ID',\n", |
||
| 165 | " 'AGE',\n", |
||
| 166 | " 'SEX',\n", |
||
| 167 | " 'ADMISSION_DATE',\n", |
||
| 168 | " 'DEPARTURE_DATE',\n", |
||
| 169 | " 'OUTCOME',\n", |
||
| 170 | " # 'DIFFICULTY_BREATHING',\n", |
||
| 171 | " # 'SUSPECT_COVID',\n", |
||
| 172 | " # 'FEVER',\n", |
||
| 173 | " # 'EMERGENCY'\n", |
||
| 174 | " ]\n", |
||
| 175 | " ]" |
||
| 176 | ] |
||
| 177 | }, |
||
| 178 | { |
||
| 179 | "cell_type": "code", |
||
| 180 | "execution_count": null, |
||
| 181 | "metadata": {}, |
||
| 182 | "outputs": [], |
||
| 183 | "source": [ |
||
| 184 | "demographic.describe().to_csv('demographic_overview.csv', mode='w', index=False)\n", |
||
| 185 | "demographic.describe()" |
||
| 186 | ] |
||
| 187 | }, |
||
| 188 | { |
||
| 189 | "cell_type": "markdown", |
||
| 190 | "metadata": {}, |
||
| 191 | "source": [ |
||
| 192 | "### Analyze data" |
||
| 193 | ] |
||
| 194 | }, |
||
| 195 | { |
||
| 196 | "cell_type": "code", |
||
| 197 | "execution_count": null, |
||
| 198 | "metadata": {}, |
||
| 199 | "outputs": [], |
||
| 200 | "source": [ |
||
| 201 | "plt.scatter(demographic['PATIENT_ID'], demographic['AGE'], s=1)\n", |
||
| 202 | "plt.xlabel('Patient Id')\n", |
||
| 203 | "plt.ylabel('Age')\n", |
||
| 204 | "plt.title('Patient-Age Scatter Plot')" |
||
| 205 | ] |
||
| 206 | }, |
||
| 207 | { |
||
| 208 | "cell_type": "code", |
||
| 209 | "execution_count": null, |
||
| 210 | "metadata": {}, |
||
| 211 | "outputs": [], |
||
| 212 | "source": [ |
||
| 213 | "plt.scatter(demographic['PATIENT_ID'], demographic['AGE'], s=1)\n", |
||
| 214 | "plt.xlabel('Patient Id')\n", |
||
| 215 | "plt.ylabel('Age')\n", |
||
| 216 | "plt.title('Patient-Age Scatter Plot')" |
||
| 217 | ] |
||
| 218 | }, |
||
| 219 | { |
||
| 220 | "cell_type": "code", |
||
| 221 | "execution_count": null, |
||
| 222 | "metadata": {}, |
||
| 223 | "outputs": [], |
||
| 224 | "source": [ |
||
| 225 | "demographic.to_csv('demographic.csv', mode='w', index=False)\n", |
||
| 226 | "demographic.head()" |
||
| 227 | ] |
||
| 228 | }, |
||
| 229 | { |
||
| 230 | "cell_type": "markdown", |
||
| 231 | "metadata": {}, |
||
| 232 | "source": [ |
||
| 233 | "## Vital Signal" |
||
| 234 | ] |
||
| 235 | }, |
||
| 236 | { |
||
| 237 | "cell_type": "code", |
||
| 238 | "execution_count": null, |
||
| 239 | "metadata": {}, |
||
| 240 | "outputs": [], |
||
| 241 | "source": [ |
||
| 242 | "vital_signs = pd.read_csv('./raw_data/19_04_2021/COVID_DSL_02.CSV', encoding='ISO-8859-1', sep='|')\n", |
||
| 243 | "print(len(vital_signs))\n", |
||
| 244 | "vital_signs.head()" |
||
| 245 | ] |
||
| 246 | }, |
||
| 247 | { |
||
| 248 | "cell_type": "code", |
||
| 249 | "execution_count": null, |
||
| 250 | "metadata": {}, |
||
| 251 | "outputs": [], |
||
| 252 | "source": [ |
||
| 253 | "vital_signs = vital_signs.rename(columns={\n", |
||
| 254 | " 'IDINGRESO': 'PATIENT_ID',\n", |
||
| 255 | " 'CONSTANTS_ING_DATE': 'RECORD_DATE',\n", |
||
| 256 | " 'CONSTANTS_ING_TIME': 'RECORD_TIME',\n", |
||
| 257 | " 'FC_HR_ING': 'HEART_RATE',\n", |
||
| 258 | " 'GLU_GLY_ING': 'BLOOD_GLUCOSE',\n", |
||
| 259 | " 'SAT_02_ING': 'OXYGEN_SATURATION',\n", |
||
| 260 | " 'TA_MAX_ING': 'MAX_BLOOD_PRESSURE',\n", |
||
| 261 | " 'TA_MIN_ING': 'MIN_BLOOD_PRESSURE',\n", |
||
| 262 | " 'TEMP_ING': 'TEMPERATURE'\n", |
||
| 263 | "})\n", |
||
| 264 | "vital_signs['RECORD_TIME'] = vital_signs['RECORD_DATE'] + ' ' + vital_signs['RECORD_TIME']\n", |
||
| 265 | "vital_signs['RECORD_TIME'] = vital_signs['RECORD_TIME'].map(lambda x: str(datetime.datetime.strptime(x, '%Y-%m-%d %H:%M')))\n", |
||
| 266 | "vital_signs = vital_signs.drop(['RECORD_DATE', 'SAT_02_ING_OBS', 'BLOOD_GLUCOSE'], axis=1)" |
||
| 267 | ] |
||
| 268 | }, |
||
| 269 | { |
||
| 270 | "cell_type": "code", |
||
| 271 | "execution_count": null, |
||
| 272 | "metadata": {}, |
||
| 273 | "outputs": [], |
||
| 274 | "source": [ |
||
| 275 | "vital_signs.describe()" |
||
| 276 | ] |
||
| 277 | }, |
||
| 278 | { |
||
| 279 | "cell_type": "code", |
||
| 280 | "execution_count": null, |
||
| 281 | "metadata": {}, |
||
| 282 | "outputs": [], |
||
| 283 | "source": [ |
||
| 284 | "vital_signs.head()" |
||
| 285 | ] |
||
| 286 | }, |
||
| 287 | { |
||
| 288 | "cell_type": "code", |
||
| 289 | "execution_count": null, |
||
| 290 | "metadata": {}, |
||
| 291 | "outputs": [], |
||
| 292 | "source": [ |
||
| 293 | "def format_temperature(x):\n", |
||
| 294 | " if type(x) == str:\n", |
||
| 295 | " return float(x.replace(',', '.'))\n", |
||
| 296 | " else:\n", |
||
| 297 | " return float(x)\n", |
||
| 298 | "\n", |
||
| 299 | "def format_oxygen(x):\n", |
||
| 300 | " x = float(x)\n", |
||
| 301 | " if x > 100:\n", |
||
| 302 | " return np.nan\n", |
||
| 303 | " else:\n", |
||
| 304 | " return x\n", |
||
| 305 | "\n", |
||
| 306 | "def format_heart_rate(x):\n", |
||
| 307 | " x = int(x)\n", |
||
| 308 | " if x > 220:\n", |
||
| 309 | " return np.nan\n", |
||
| 310 | " else:\n", |
||
| 311 | " return x\n", |
||
| 312 | "\n", |
||
| 313 | "vital_signs['TEMPERATURE'] = vital_signs['TEMPERATURE'].map(lambda x: format_temperature(x))\n", |
||
| 314 | "vital_signs['OXYGEN_SATURATION'] = vital_signs['OXYGEN_SATURATION'].map(lambda x: format_oxygen(x))\n", |
||
| 315 | "vital_signs['HEART_RATE'] = vital_signs['HEART_RATE'].map(lambda x: format_heart_rate(x))" |
||
| 316 | ] |
||
| 317 | }, |
||
| 318 | { |
||
| 319 | "cell_type": "code", |
||
| 320 | "execution_count": null, |
||
| 321 | "metadata": {}, |
||
| 322 | "outputs": [], |
||
| 323 | "source": [ |
||
| 324 | "vital_signs = vital_signs.replace(0, np.NAN)" |
||
| 325 | ] |
||
| 326 | }, |
||
| 327 | { |
||
| 328 | "cell_type": "code", |
||
| 329 | "execution_count": null, |
||
| 330 | "metadata": {}, |
||
| 331 | "outputs": [], |
||
| 332 | "source": [ |
||
| 333 | "vital_signs = vital_signs.groupby(['PATIENT_ID', 'RECORD_TIME'], dropna=True, as_index = False).mean()\n", |
||
| 334 | "vital_signs.head()" |
||
| 335 | ] |
||
| 336 | }, |
||
| 337 | { |
||
| 338 | "cell_type": "code", |
||
| 339 | "execution_count": null, |
||
| 340 | "metadata": {}, |
||
| 341 | "outputs": [], |
||
| 342 | "source": [ |
||
| 343 | "vital_signs.describe()" |
||
| 344 | ] |
||
| 345 | }, |
||
| 346 | { |
||
| 347 | "cell_type": "code", |
||
| 348 | "execution_count": null, |
||
| 349 | "metadata": {}, |
||
| 350 | "outputs": [], |
||
| 351 | "source": [ |
||
| 352 | "vital_signs.describe().to_csv('vital_signs_overview.csv', index=False, mode='w')\n", |
||
| 353 | "vital_signs.describe()" |
||
| 354 | ] |
||
| 355 | }, |
||
| 356 | { |
||
| 357 | "cell_type": "code", |
||
| 358 | "execution_count": null, |
||
| 359 | "metadata": {}, |
||
| 360 | "outputs": [], |
||
| 361 | "source": [ |
||
| 362 | "#plt.rcParams['figure.figsize'] = [10, 5]\n", |
||
| 363 | "fig=plt.figure(figsize=(16,10), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 364 | "\n", |
||
| 365 | "plt.subplot(2, 3, 1)\n", |
||
| 366 | "plt.scatter(vital_signs.index, vital_signs['MAX_BLOOD_PRESSURE'], s=1)\n", |
||
| 367 | "plt.xlabel('Index')\n", |
||
| 368 | "plt.ylabel('Max Blood Pressure')\n", |
||
| 369 | "plt.title('Visit-Max Blood Pressure Scatter Plot')\n", |
||
| 370 | "\n", |
||
| 371 | "plt.subplot(2, 3, 2)\n", |
||
| 372 | "plt.scatter(vital_signs.index, vital_signs['MIN_BLOOD_PRESSURE'], s=1)\n", |
||
| 373 | "plt.xlabel('Index')\n", |
||
| 374 | "plt.ylabel('Min Blood Pressure')\n", |
||
| 375 | "plt.title('Visit-Min Blood Pressure Scatter Plot')\n", |
||
| 376 | "\n", |
||
| 377 | "plt.subplot(2, 3, 3)\n", |
||
| 378 | "plt.scatter(vital_signs.index, vital_signs['TEMPERATURE'], s=1)\n", |
||
| 379 | "plt.xlabel('Index')\n", |
||
| 380 | "plt.ylabel('Temperature')\n", |
||
| 381 | "plt.title('Visit-Temperature Scatter Plot')\n", |
||
| 382 | "\n", |
||
| 383 | "plt.subplot(2, 3, 4)\n", |
||
| 384 | "plt.scatter(vital_signs.index, vital_signs['HEART_RATE'], s=1)\n", |
||
| 385 | "plt.xlabel('Index')\n", |
||
| 386 | "plt.ylabel('Heart Rate')\n", |
||
| 387 | "plt.title('Visit-Heart Rate Scatter Plot')\n", |
||
| 388 | "\n", |
||
| 389 | "plt.subplot(2, 3, 5)\n", |
||
| 390 | "plt.scatter(vital_signs.index, vital_signs['OXYGEN_SATURATION'], s=1)\n", |
||
| 391 | "plt.xlabel('Index')\n", |
||
| 392 | "plt.ylabel('Oxygen Saturation')\n", |
||
| 393 | "plt.title('Visit-Oxygen Saturation Scatter Plot')\n", |
||
| 394 | "\n", |
||
| 395 | "plt.show()" |
||
| 396 | ] |
||
| 397 | }, |
||
| 398 | { |
||
| 399 | "cell_type": "code", |
||
| 400 | "execution_count": null, |
||
| 401 | "metadata": {}, |
||
| 402 | "outputs": [], |
||
| 403 | "source": [ |
||
| 404 | "#plt.rcParams['figure.figsize'] = [10, 5]\n", |
||
| 405 | "fig=plt.figure(figsize=(16,10), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 406 | "\n", |
||
| 407 | "plt.subplot(2, 3, 1)\n", |
||
| 408 | "plt.hist(vital_signs['MAX_BLOOD_PRESSURE'], bins=30)\n", |
||
| 409 | "plt.xlabel('Index')\n", |
||
| 410 | "plt.ylabel('Max Blood Pressure')\n", |
||
| 411 | "plt.title('Visit-Max Blood Pressure Histogram')\n", |
||
| 412 | "\n", |
||
| 413 | "plt.subplot(2, 3, 2)\n", |
||
| 414 | "plt.hist(vital_signs['MIN_BLOOD_PRESSURE'], bins=30)\n", |
||
| 415 | "plt.xlabel('Index')\n", |
||
| 416 | "plt.ylabel('Min Blood Pressure')\n", |
||
| 417 | "plt.title('Visit-Min Blood Pressure Histogram')\n", |
||
| 418 | "\n", |
||
| 419 | "plt.subplot(2, 3, 3)\n", |
||
| 420 | "plt.hist(vital_signs['TEMPERATURE'], bins=30)\n", |
||
| 421 | "plt.xlabel('Index')\n", |
||
| 422 | "plt.ylabel('Temperature')\n", |
||
| 423 | "plt.title('Visit-Temperature Histogram')\n", |
||
| 424 | "\n", |
||
| 425 | "plt.subplot(2, 3, 4)\n", |
||
| 426 | "plt.hist(vital_signs['HEART_RATE'], bins=30)\n", |
||
| 427 | "plt.xlabel('Index')\n", |
||
| 428 | "plt.ylabel('Heart Rate')\n", |
||
| 429 | "plt.title('Visit-Heart Rate Histogram')\n", |
||
| 430 | "\n", |
||
| 431 | "plt.subplot(2, 3, 5)\n", |
||
| 432 | "plt.hist(vital_signs['OXYGEN_SATURATION'], bins=30)\n", |
||
| 433 | "plt.xlabel('Index')\n", |
||
| 434 | "plt.ylabel('Oxygen Saturation')\n", |
||
| 435 | "plt.title('Visit-Oxygen Saturation Histogram')\n", |
||
| 436 | "\n", |
||
| 437 | "plt.show()" |
||
| 438 | ] |
||
| 439 | }, |
||
| 440 | { |
||
| 441 | "cell_type": "markdown", |
||
| 442 | "metadata": {}, |
||
| 443 | "source": [ |
||
| 444 | "### Missing rate of each visit" |
||
| 445 | ] |
||
| 446 | }, |
||
| 447 | { |
||
| 448 | "cell_type": "code", |
||
| 449 | "execution_count": null, |
||
| 450 | "metadata": {}, |
||
| 451 | "outputs": [], |
||
| 452 | "source": [ |
||
| 453 | "sum(vital_signs.T.isnull().sum()) / ((len(vital_signs.T) - 2) * len(vital_signs))" |
||
| 454 | ] |
||
| 455 | }, |
||
| 456 | { |
||
| 457 | "cell_type": "markdown", |
||
| 458 | "metadata": {}, |
||
| 459 | "source": [ |
||
| 460 | "### Normalize data" |
||
| 461 | ] |
||
| 462 | }, |
||
| 463 | { |
||
| 464 | "cell_type": "code", |
||
| 465 | "execution_count": null, |
||
| 466 | "metadata": {}, |
||
| 467 | "outputs": [], |
||
| 468 | "source": [ |
||
| 469 | "\"\"\"\n", |
||
| 470 | "for key in vital_signs.keys()[2:]:\n", |
||
| 471 | " vital_signs[key] = (vital_signs[key] - vital_signs[key].mean()) / (vital_signs[key].std() + 1e-12)\n", |
||
| 472 | "\n", |
||
| 473 | "vital_signs.describe()\n", |
||
| 474 | "\"\"\"" |
||
| 475 | ] |
||
| 476 | }, |
||
| 477 | { |
||
| 478 | "cell_type": "code", |
||
| 479 | "execution_count": null, |
||
| 480 | "metadata": {}, |
||
| 481 | "outputs": [], |
||
| 482 | "source": [ |
||
| 483 | "vital_signs.to_csv('visual_signs.csv', mode='w', index=False)" |
||
| 484 | ] |
||
| 485 | }, |
||
| 486 | { |
||
| 487 | "cell_type": "code", |
||
| 488 | "execution_count": null, |
||
| 489 | "metadata": {}, |
||
| 490 | "outputs": [], |
||
| 491 | "source": [ |
||
| 492 | "len(vital_signs) / len(vital_signs['PATIENT_ID'].unique())" |
||
| 493 | ] |
||
| 494 | }, |
||
| 495 | { |
||
| 496 | "cell_type": "markdown", |
||
| 497 | "metadata": {}, |
||
| 498 | "source": [ |
||
| 499 | "## Lab Tests" |
||
| 500 | ] |
||
| 501 | }, |
||
| 502 | { |
||
| 503 | "cell_type": "code", |
||
| 504 | "execution_count": null, |
||
| 505 | "metadata": {}, |
||
| 506 | "outputs": [], |
||
| 507 | "source": [ |
||
| 508 | "lab_tests = pd.read_csv('./raw_data/19_04_2021/COVID_DSL_06_v2.CSV', encoding='ISO-8859-1', sep=';')\n", |
||
| 509 | "lab_tests = lab_tests.rename(columns={'IDINGRESO': 'PATIENT_ID'})\n", |
||
| 510 | "print(len(lab_tests))\n", |
||
| 511 | "\n", |
||
| 512 | "# del useless data\n", |
||
| 513 | "lab_tests = lab_tests[\n", |
||
| 514 | " [\n", |
||
| 515 | " 'PATIENT_ID',\n", |
||
| 516 | " 'LAB_NUMBER',\n", |
||
| 517 | " 'LAB_DATE',\n", |
||
| 518 | " 'TIME_LAB',\n", |
||
| 519 | " 'ITEM_LAB',\n", |
||
| 520 | " 'VAL_RESULT'\n", |
||
| 521 | " # UD_RESULT: unit\n", |
||
| 522 | " # REF_VALUES: reference values\n", |
||
| 523 | " ]\n", |
||
| 524 | " ]\n", |
||
| 525 | "\n", |
||
| 526 | "lab_tests.head()" |
||
| 527 | ] |
||
| 528 | }, |
||
| 529 | { |
||
| 530 | "cell_type": "code", |
||
| 531 | "execution_count": null, |
||
| 532 | "metadata": {}, |
||
| 533 | "outputs": [], |
||
| 534 | "source": [ |
||
| 535 | "lab_tests = lab_tests.groupby(['PATIENT_ID', 'LAB_NUMBER', 'LAB_DATE', 'TIME_LAB', 'ITEM_LAB'], dropna=True, as_index = False).first()\n", |
||
| 536 | "lab_tests = lab_tests.set_index(['PATIENT_ID', 'LAB_NUMBER', 'LAB_DATE', 'TIME_LAB', 'ITEM_LAB'], drop = True).unstack('ITEM_LAB')['VAL_RESULT'].reset_index()\n", |
||
| 537 | "\n", |
||
| 538 | "lab_tests = lab_tests.drop([\n", |
||
| 539 | " 'CFLAG -- ALARMA HEMOGRAMA', \n", |
||
| 540 | " 'CORONA -- PCR CORONAVIRUS 2019nCoV', \n", |
||
| 541 | " 'CRIOGLO -- CRIOGLOBULINAS',\n", |
||
| 542 | " 'EGCOVID -- ESTUDIO GENETICO COVID-19',\n", |
||
| 543 | " 'FRO1 -- ',\n", |
||
| 544 | " 'FRO1 -- FROTIS EN SANGRE PERIFERICA',\n", |
||
| 545 | " 'FRO2 -- ',\n", |
||
| 546 | " 'FRO2 -- FROTIS EN SANGRE PERIFERICA',\n", |
||
| 547 | " 'FRO3 -- ',\n", |
||
| 548 | " 'FRO3 -- FROTIS EN SANGRE PERIFERICA',\n", |
||
| 549 | " 'FRO_COMEN -- ',\n", |
||
| 550 | " 'FRO_COMEN -- FROTIS EN SANGRE PERIFERICA',\n", |
||
| 551 | " 'G-CORONAV (RT-PCR) -- Tipo de muestra: ASPIRADO BRONCOALVEOLAR',\n", |
||
| 552 | " 'G-CORONAV (RT-PCR) -- Tipo de muestra: EXUDADO',\n", |
||
| 553 | " 'GRRH -- GRUPO SANGUÖNEO Y FACTOR Rh',\n", |
||
| 554 | " 'HEML -- RECUENTO CELULAR LIQUIDO',\n", |
||
| 555 | " 'HEML -- Recuento Hemat¡es',\n", |
||
| 556 | " 'IFSUERO -- INMUNOFIJACION EN SUERO',\n", |
||
| 557 | " 'OBS_BIOMOL -- OBSERVACIONES GENETICA MOLECULAR',\n", |
||
| 558 | " 'OBS_BIOO -- Observaciones Bioqu¡mica Orina',\n", |
||
| 559 | " 'OBS_CB -- Observaciones Coagulaci¢n',\n", |
||
| 560 | " 'OBS_GASES -- Observaciones Gasometr¡a Arterial',\n", |
||
| 561 | " 'OBS_GASV -- Observaciones Gasometr¡a Venosa',\n", |
||
| 562 | " 'OBS_GEN2 -- OBSERVACIONES GENETICA',\n", |
||
| 563 | " 'OBS_HOR -- Observaciones Hormonas',\n", |
||
| 564 | " 'OBS_MICRO -- Observaciones Microbiolog¡a',\n", |
||
| 565 | " 'OBS_NULA2 -- Observaciones Bioqu¡mica',\n", |
||
| 566 | " 'OBS_NULA3 -- Observaciones Hematolog¡a',\n", |
||
| 567 | " 'OBS_PESP -- Observaciones Pruebas especiales',\n", |
||
| 568 | " 'OBS_SERO -- Observaciones Serolog¡a',\n", |
||
| 569 | " 'OBS_SIS -- Observaciones Orina',\n", |
||
| 570 | " 'PCR VIRUS RESPIRATORIOS -- Tipo de muestra: ASPIRADO BRONCOALVEOLAR',\n", |
||
| 571 | " 'PCR VIRUS RESPIRATORIOS -- Tipo de muestra: BAS',\n", |
||
| 572 | " 'PCR VIRUS RESPIRATORIOS -- Tipo de muestra: ESPUTO',\n", |
||
| 573 | " 'PCR VIRUS RESPIRATORIOS -- Tipo de muestra: EXUDADO',\n", |
||
| 574 | " 'PCR VIRUS RESPIRATORIOS -- Tipo de muestra: LAVADO BRONCOALVEOLAR',\n", |
||
| 575 | " 'PCR VIRUS RESPIRATORIOS -- Tipo de muestra: LAVADO NASOFARÖNGEO',\n", |
||
| 576 | " 'PTGOR -- PROTEINOGRAMA ORINA',\n", |
||
| 577 | " 'RESUL_IFT -- ESTUDIO DE INMUNOFENOTIPO',\n", |
||
| 578 | " 'RESUL_IFT -- Resultado',\n", |
||
| 579 | " 'Resultado -- Resultado',\n", |
||
| 580 | " 'SED1 -- ',\n", |
||
| 581 | " 'SED1 -- SEDIMENTO',\n", |
||
| 582 | " 'SED2 -- ',\n", |
||
| 583 | " 'SED2 -- SEDIMENTO',\n", |
||
| 584 | " 'SED3 -- ',\n", |
||
| 585 | " 'SED3 -- SEDIMENTO',\n", |
||
| 586 | " 'TIPOL -- TIPO DE LIQUIDO',\n", |
||
| 587 | " 'Tecnica -- T\\x82cnica',\n", |
||
| 588 | " 'TpMues -- Tipo de muestra',\n", |
||
| 589 | " 'VHCBLOT -- INMUNOBLOT VIRUS HEPATITIS C',\n", |
||
| 590 | " 'VIR_TM -- VIRUS TIPO DE MUESTRA',\n", |
||
| 591 | " 'LEGIORI -- AG. LEGIONELA PNEUMOPHILA EN ORINA',\n", |
||
| 592 | " 'NEUMOORI -- AG NEUMOCOCO EN ORINA',\n", |
||
| 593 | " 'VIHAC -- VIH AC'\n", |
||
| 594 | " ], axis=1)\n", |
||
| 595 | "\n", |
||
| 596 | " \n", |
||
| 597 | "lab_tests.head()" |
||
| 598 | ] |
||
| 599 | }, |
||
| 600 | { |
||
| 601 | "cell_type": "code", |
||
| 602 | "execution_count": null, |
||
| 603 | "metadata": {}, |
||
| 604 | "outputs": [], |
||
| 605 | "source": [ |
||
| 606 | "lab_tests = lab_tests.replace('Sin resultado.', np.nan)\n", |
||
| 607 | "lab_tests = lab_tests.replace('Sin resultado', np.nan)\n", |
||
| 608 | "lab_tests = lab_tests.replace('----', np.nan).replace('---', np.nan)\n", |
||
| 609 | "lab_tests = lab_tests.replace('> ', '').replace('< ', '')\n", |
||
| 610 | "\n", |
||
| 611 | "def change_format(x):\n", |
||
| 612 | " if x is None:\n", |
||
| 613 | " return np.nan\n", |
||
| 614 | " elif type(x) == str:\n", |
||
| 615 | " if x.startswith('Negativo ('):\n", |
||
| 616 | " return x.replace('Negativo (', '-')[:-1]\n", |
||
| 617 | " elif x.startswith('Positivo ('):\n", |
||
| 618 | " return x.replace('Positivo (', '')[:-1]\n", |
||
| 619 | " elif x.startswith('Zona limite ('):\n", |
||
| 620 | " return x.replace('Zona limite (', '')[:-1]\n", |
||
| 621 | " elif x.startswith('>'):\n", |
||
| 622 | " return x.replace('> ', '').replace('>', '')\n", |
||
| 623 | " elif x.startswith('<'):\n", |
||
| 624 | " return x.replace('< ', '').replace('<', '')\n", |
||
| 625 | " elif x.endswith(' mg/dl'):\n", |
||
| 626 | " return x.replace(' mg/dl', '')\n", |
||
| 627 | " elif x.endswith('/æl'):\n", |
||
| 628 | " return x.replace('/æl', '')\n", |
||
| 629 | " elif x.endswith(' copias/mL'):\n", |
||
| 630 | " return x.replace(' copias/mL', '')\n", |
||
| 631 | " elif x == 'Numerosos':\n", |
||
| 632 | " return 1.5\n", |
||
| 633 | " elif x == 'Aislados':\n", |
||
| 634 | " return 0.5\n", |
||
| 635 | " elif x == 'Se detecta' or x == 'Se observan' or x == 'Normal' or x == 'Positivo':\n", |
||
| 636 | " return 1\n", |
||
| 637 | " elif x == 'No se detecta' or x == 'No se observan' or x == 'Negativo':\n", |
||
| 638 | " return 0\n", |
||
| 639 | " elif x == 'Indeterminado':\n", |
||
| 640 | " return np.nan\n", |
||
| 641 | " else:\n", |
||
| 642 | " num = re.findall(\"[-+]?\\d+\\.\\d+\", x)\n", |
||
| 643 | " if len(num) == 0:\n", |
||
| 644 | " return np.nan\n", |
||
| 645 | " else:\n", |
||
| 646 | " return num[0]\n", |
||
| 647 | " else:\n", |
||
| 648 | " return x\n", |
||
| 649 | "\n", |
||
| 650 | "feature_value_dict = dict()\n", |
||
| 651 | "\n", |
||
| 652 | "for k in tqdm(lab_tests.keys()[4:]):\n", |
||
| 653 | " lab_tests[k] = lab_tests[k].map(lambda x: change_format(change_format(x)))\n", |
||
| 654 | " feature_value_dict[k] = lab_tests[k].unique()" |
||
| 655 | ] |
||
| 656 | }, |
||
| 657 | { |
||
| 658 | "cell_type": "code", |
||
| 659 | "execution_count": null, |
||
| 660 | "metadata": {}, |
||
| 661 | "outputs": [], |
||
| 662 | "source": [ |
||
| 663 | "def nan_and_not_nan(x):\n", |
||
| 664 | " if x == x:\n", |
||
| 665 | " return 1\n", |
||
| 666 | " else: # nan\n", |
||
| 667 | " return 0\n", |
||
| 668 | "\n", |
||
| 669 | "def is_float(num):\n", |
||
| 670 | " try:\n", |
||
| 671 | " float(num)\n", |
||
| 672 | " return True\n", |
||
| 673 | " except ValueError:\n", |
||
| 674 | " return False\n", |
||
| 675 | "\n", |
||
| 676 | "def is_all_float(x):\n", |
||
| 677 | " for i in x:\n", |
||
| 678 | " if i == i and (i != None):\n", |
||
| 679 | " if not is_float(i):\n", |
||
| 680 | " return False\n", |
||
| 681 | " return True\n", |
||
| 682 | "\n", |
||
| 683 | "def to_float(x):\n", |
||
| 684 | " if x != None:\n", |
||
| 685 | " return float(x)\n", |
||
| 686 | " else:\n", |
||
| 687 | " return np.nan\n", |
||
| 688 | "\n", |
||
| 689 | "other_feature_dict = dict()\n", |
||
| 690 | "\n", |
||
| 691 | "for feature in tqdm(feature_value_dict.keys()):\n", |
||
| 692 | " values = feature_value_dict[feature]\n", |
||
| 693 | " if is_all_float(values):\n", |
||
| 694 | " lab_tests[feature] = lab_tests[feature].map(lambda x: to_float(x))\n", |
||
| 695 | " elif len(values) == 2:\n", |
||
| 696 | " lab_tests[feature] = lab_tests[feature].map(lambda x: nan_and_not_nan(x))\n", |
||
| 697 | " else:\n", |
||
| 698 | " other_feature_dict[feature] = values" |
||
| 699 | ] |
||
| 700 | }, |
||
| 701 | { |
||
| 702 | "cell_type": "code", |
||
| 703 | "execution_count": null, |
||
| 704 | "metadata": {}, |
||
| 705 | "outputs": [], |
||
| 706 | "source": [ |
||
| 707 | "other_feature_dict" |
||
| 708 | ] |
||
| 709 | }, |
||
| 710 | { |
||
| 711 | "cell_type": "code", |
||
| 712 | "execution_count": null, |
||
| 713 | "metadata": {}, |
||
| 714 | "outputs": [], |
||
| 715 | "source": [ |
||
| 716 | "def format_time(t):\n", |
||
| 717 | " if '/' in t:\n", |
||
| 718 | " return str(datetime.datetime.strptime(t, '%d/%m/%Y %H:%M'))\n", |
||
| 719 | " else:\n", |
||
| 720 | " return str(datetime.datetime.strptime(t, '%d-%m-%Y %H:%M'))\n", |
||
| 721 | "\n", |
||
| 722 | "lab_tests['RECORD_TIME'] = lab_tests['LAB_DATE'] + ' ' + lab_tests['TIME_LAB']\n", |
||
| 723 | "lab_tests['RECORD_TIME'] = lab_tests['RECORD_TIME'].map(lambda x: format_time(x))\n", |
||
| 724 | "lab_tests = lab_tests.drop(['LAB_NUMBER', 'LAB_DATE', 'TIME_LAB'], axis=1)\n", |
||
| 725 | "# lab_tests = lab_tests.drop(['LAB_NUMBER', 'TIME_LAB'], axis=1)\n", |
||
| 726 | "lab_tests.head()" |
||
| 727 | ] |
||
| 728 | }, |
||
| 729 | { |
||
| 730 | "cell_type": "code", |
||
| 731 | "execution_count": null, |
||
| 732 | "metadata": {}, |
||
| 733 | "outputs": [], |
||
| 734 | "source": [ |
||
| 735 | "lab_tests_patient = lab_tests.groupby(['PATIENT_ID'], dropna=True, as_index = False).mean()\n", |
||
| 736 | "print(len(lab_tests_patient))\n", |
||
| 737 | "count = [i for i in lab_tests_patient.count()[1:]]\n", |
||
| 738 | "plt.hist(count)" |
||
| 739 | ] |
||
| 740 | }, |
||
| 741 | { |
||
| 742 | "cell_type": "code", |
||
| 743 | "execution_count": null, |
||
| 744 | "metadata": {}, |
||
| 745 | "outputs": [], |
||
| 746 | "source": [ |
||
| 747 | "patient_total = len(lab_tests_patient)\n", |
||
| 748 | "threshold = patient_total * 0.1\n", |
||
| 749 | "reserved_keys = []\n", |
||
| 750 | "\n", |
||
| 751 | "for key in lab_tests_patient.keys():\n", |
||
| 752 | " if lab_tests_patient[key].count() > threshold:\n", |
||
| 753 | " reserved_keys.append(key)\n", |
||
| 754 | "\n", |
||
| 755 | "print(len(reserved_keys))\n", |
||
| 756 | "reserved_keys" |
||
| 757 | ] |
||
| 758 | }, |
||
| 759 | { |
||
| 760 | "cell_type": "code", |
||
| 761 | "execution_count": null, |
||
| 762 | "metadata": {}, |
||
| 763 | "outputs": [], |
||
| 764 | "source": [ |
||
| 765 | "reserved_keys.insert(1, 'RECORD_TIME')\n", |
||
| 766 | "\n", |
||
| 767 | "lab_tests = lab_tests.groupby(['PATIENT_ID', 'RECORD_TIME'], dropna=True, as_index = False).mean()\n", |
||
| 768 | "\n", |
||
| 769 | "lab_tests = lab_tests[reserved_keys]\n", |
||
| 770 | "lab_tests.head()" |
||
| 771 | ] |
||
| 772 | }, |
||
| 773 | { |
||
| 774 | "cell_type": "markdown", |
||
| 775 | "metadata": {}, |
||
| 776 | "source": [ |
||
| 777 | "### Missing rate of each visit" |
||
| 778 | ] |
||
| 779 | }, |
||
| 780 | { |
||
| 781 | "cell_type": "code", |
||
| 782 | "execution_count": null, |
||
| 783 | "metadata": {}, |
||
| 784 | "outputs": [], |
||
| 785 | "source": [ |
||
| 786 | "sum(lab_tests.T.isnull().sum()) / ((len(lab_tests.T) - 2) * len(lab_tests))" |
||
| 787 | ] |
||
| 788 | }, |
||
| 789 | { |
||
| 790 | "cell_type": "markdown", |
||
| 791 | "metadata": {}, |
||
| 792 | "source": [ |
||
| 793 | "### Scatter Plot" |
||
| 794 | ] |
||
| 795 | }, |
||
| 796 | { |
||
| 797 | "cell_type": "code", |
||
| 798 | "execution_count": null, |
||
| 799 | "metadata": {}, |
||
| 800 | "outputs": [], |
||
| 801 | "source": [ |
||
| 802 | "fig=plt.figure(figsize=(16,200), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 803 | "\n", |
||
| 804 | "i = 1\n", |
||
| 805 | "for key in lab_tests.keys()[2:]:\n", |
||
| 806 | " plt.subplot(33, 3, i)\n", |
||
| 807 | " plt.scatter(lab_tests.index, lab_tests[key], s=1)\n", |
||
| 808 | " plt.ylabel(key)\n", |
||
| 809 | " i += 1\n", |
||
| 810 | "\n", |
||
| 811 | "plt.show()" |
||
| 812 | ] |
||
| 813 | }, |
||
| 814 | { |
||
| 815 | "cell_type": "code", |
||
| 816 | "execution_count": null, |
||
| 817 | "metadata": {}, |
||
| 818 | "outputs": [], |
||
| 819 | "source": [ |
||
| 820 | "fig=plt.figure(figsize=(20,120), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 821 | "\n", |
||
| 822 | "i = 1\n", |
||
| 823 | "for key in lab_tests.keys()[2:]:\n", |
||
| 824 | " plt.subplot(23, 4, i)\n", |
||
| 825 | " plt.hist(lab_tests[key], bins=30)\n", |
||
| 826 | " q3 = lab_tests[key].quantile(0.75)\n", |
||
| 827 | " q1 = lab_tests[key].quantile(0.25)\n", |
||
| 828 | " qh = q3 + 3 * (q3 - q1)\n", |
||
| 829 | " ql = q1 - 3 * (q3 - q1)\n", |
||
| 830 | " sigma = 5\n", |
||
| 831 | " plt.axline([sigma*lab_tests[key].std() + lab_tests[key].mean(), 0], [sigma*lab_tests[key].std() + lab_tests[key].mean(), 1], color = \"r\", linestyle=(0, (5, 5)))\n", |
||
| 832 | " plt.axline([-sigma*lab_tests[key].std() + lab_tests[key].mean(), 0], [-sigma*lab_tests[key].std() + lab_tests[key].mean(), 1], color = \"r\", linestyle=(0, (5, 5)))\n", |
||
| 833 | " #plt.axline([lab_tests[key].quantile(0.25), 0], [lab_tests[key].quantile(0.25), 1], color = \"k\", linestyle=(0, (5, 5)))\n", |
||
| 834 | " #plt.axline([lab_tests[key].quantile(0.75), 0], [lab_tests[key].quantile(0.75), 1], color = \"k\", linestyle=(0, (5, 5)))\n", |
||
| 835 | " plt.axline([qh, 0], [qh, 1], color='k', linestyle=(0, (5, 5)))\n", |
||
| 836 | " plt.axline([ql, 0], [ql, 1], color='k', linestyle=(0, (5, 5)))\n", |
||
| 837 | " plt.ylabel(key)\n", |
||
| 838 | " i += 1\n", |
||
| 839 | "\n", |
||
| 840 | "plt.show()" |
||
| 841 | ] |
||
| 842 | }, |
||
| 843 | { |
||
| 844 | "cell_type": "markdown", |
||
| 845 | "metadata": {}, |
||
| 846 | "source": [ |
||
| 847 | "### Normalize data" |
||
| 848 | ] |
||
| 849 | }, |
||
| 850 | { |
||
| 851 | "cell_type": "code", |
||
| 852 | "execution_count": null, |
||
| 853 | "metadata": {}, |
||
| 854 | "outputs": [], |
||
| 855 | "source": [ |
||
| 856 | "\"\"\"\n", |
||
| 857 | "for key in lab_tests.keys()[2:]:\n", |
||
| 858 | " lab_tests[key] = (lab_tests[key] - lab_tests[key].mean()) / (lab_tests[key].std() + 1e-12)\n", |
||
| 859 | "\n", |
||
| 860 | "lab_tests.describe()\n", |
||
| 861 | "\"\"\"" |
||
| 862 | ] |
||
| 863 | }, |
||
| 864 | { |
||
| 865 | "cell_type": "code", |
||
| 866 | "execution_count": null, |
||
| 867 | "metadata": {}, |
||
| 868 | "outputs": [], |
||
| 869 | "source": [ |
||
| 870 | "# 【del normalization】\n", |
||
| 871 | "# for key in lab_tests.keys()[2:]:\n", |
||
| 872 | "# r = lab_tests[lab_tests[key].between(lab_tests[key].quantile(0.05), lab_tests[key].quantile(0.95))]\n", |
||
| 873 | "# lab_tests[key] = (lab_tests[key] - r[key].mean()) / (r[key].std() + 1e-12)" |
||
| 874 | ] |
||
| 875 | }, |
||
| 876 | { |
||
| 877 | "cell_type": "code", |
||
| 878 | "execution_count": null, |
||
| 879 | "metadata": {}, |
||
| 880 | "outputs": [], |
||
| 881 | "source": [ |
||
| 882 | "lab_tests.to_csv('lab_test.csv', mode='w', index=False)" |
||
| 883 | ] |
||
| 884 | }, |
||
| 885 | { |
||
| 886 | "cell_type": "markdown", |
||
| 887 | "metadata": {}, |
||
| 888 | "source": [ |
||
| 889 | "# Concat data" |
||
| 890 | ] |
||
| 891 | }, |
||
| 892 | { |
||
| 893 | "cell_type": "code", |
||
| 894 | "execution_count": null, |
||
| 895 | "metadata": {}, |
||
| 896 | "outputs": [], |
||
| 897 | "source": [ |
||
| 898 | "demographic['PATIENT_ID'] = demographic['PATIENT_ID'].map(lambda x: str(int(x)))\n", |
||
| 899 | "vital_signs['PATIENT_ID'] = vital_signs['PATIENT_ID'].map(lambda x: str(int(x)))\n", |
||
| 900 | "lab_tests['PATIENT_ID'] = lab_tests['PATIENT_ID'].map(lambda x: str(int(x)))" |
||
| 901 | ] |
||
| 902 | }, |
||
| 903 | { |
||
| 904 | "cell_type": "code", |
||
| 905 | "execution_count": null, |
||
| 906 | "metadata": {}, |
||
| 907 | "outputs": [], |
||
| 908 | "source": [ |
||
| 909 | "len(demographic['PATIENT_ID'].unique()), len(vital_signs['PATIENT_ID'].unique()), len(lab_tests['PATIENT_ID'].unique())" |
||
| 910 | ] |
||
| 911 | }, |
||
| 912 | { |
||
| 913 | "cell_type": "code", |
||
| 914 | "execution_count": null, |
||
| 915 | "metadata": {}, |
||
| 916 | "outputs": [], |
||
| 917 | "source": [ |
||
| 918 | "train_df = pd.merge(vital_signs, lab_tests, on=['PATIENT_ID', 'RECORD_TIME'], how='outer')\n", |
||
| 919 | "\n", |
||
| 920 | "train_df = train_df.groupby(['PATIENT_ID', 'RECORD_TIME'], dropna=True, as_index = False).mean()\n", |
||
| 921 | "\n", |
||
| 922 | "train_df = pd.merge(demographic, train_df, on=['PATIENT_ID'], how='left')\n", |
||
| 923 | "\n", |
||
| 924 | "train_df.head()" |
||
| 925 | ] |
||
| 926 | }, |
||
| 927 | { |
||
| 928 | "cell_type": "code", |
||
| 929 | "execution_count": null, |
||
| 930 | "metadata": {}, |
||
| 931 | "outputs": [], |
||
| 932 | "source": [ |
||
| 933 | "# del rows without patient_id, admission_date, record_time, or outcome\n", |
||
| 934 | "train_df = train_df.dropna(axis=0, how='any', subset=['PATIENT_ID', 'ADMISSION_DATE', 'RECORD_TIME', 'OUTCOME'])" |
||
| 935 | ] |
||
| 936 | }, |
||
| 937 | { |
||
| 938 | "cell_type": "code", |
||
| 939 | "execution_count": null, |
||
| 940 | "metadata": {}, |
||
| 941 | "outputs": [], |
||
| 942 | "source": [ |
||
| 943 | "train_df.to_csv('train.csv', mode='w', index=False)\n", |
||
| 944 | "train_df.describe()" |
||
| 945 | ] |
||
| 946 | }, |
||
| 947 | { |
||
| 948 | "cell_type": "markdown", |
||
| 949 | "metadata": {}, |
||
| 950 | "source": [ |
||
| 951 | "## Missing rate of each visit" |
||
| 952 | ] |
||
| 953 | }, |
||
| 954 | { |
||
| 955 | "cell_type": "code", |
||
| 956 | "execution_count": null, |
||
| 957 | "metadata": {}, |
||
| 958 | "outputs": [], |
||
| 959 | "source": [ |
||
| 960 | "sum(train_df.T.isnull().sum()) / ((len(train_df.T) - 2) * len(train_df))" |
||
| 961 | ] |
||
| 962 | }, |
||
| 963 | { |
||
| 964 | "cell_type": "markdown", |
||
| 965 | "metadata": {}, |
||
| 966 | "source": [ |
||
| 967 | "# Split and save data" |
||
| 968 | ] |
||
| 969 | }, |
||
| 970 | { |
||
| 971 | "cell_type": "markdown", |
||
| 972 | "metadata": {}, |
||
| 973 | "source": [ |
||
| 974 | "* demo: demographic data\n", |
||
| 975 | "* x: lab test & vital signs\n", |
||
| 976 | "* y: outcome & length of stay" |
||
| 977 | ] |
||
| 978 | }, |
||
| 979 | { |
||
| 980 | "cell_type": "code", |
||
| 981 | "execution_count": null, |
||
| 982 | "metadata": {}, |
||
| 983 | "outputs": [], |
||
| 984 | "source": [ |
||
| 985 | "patient_ids = train_df['PATIENT_ID'].unique()\n", |
||
| 986 | "\n", |
||
| 987 | "demo_cols = ['AGE', 'SEX'] # , 'DIFFICULTY_BREATHING', 'FEVER', 'SUSPECT_COVID', 'EMERGENCY'\n", |
||
| 988 | "test_cols = []\n", |
||
| 989 | "\n", |
||
| 990 | "# get column names\n", |
||
| 991 | "for k in train_df.keys():\n", |
||
| 992 | " if not k in demographic.keys():\n", |
||
| 993 | " if not k == 'RECORD_TIME':\n", |
||
| 994 | " test_cols.append(k)\n", |
||
| 995 | "\n", |
||
| 996 | "test_median = train_df[test_cols].median()" |
||
| 997 | ] |
||
| 998 | }, |
||
| 999 | { |
||
| 1000 | "cell_type": "code", |
||
| 1001 | "execution_count": null, |
||
| 1002 | "metadata": {}, |
||
| 1003 | "outputs": [], |
||
| 1004 | "source": [ |
||
| 1005 | "test_cols" |
||
| 1006 | ] |
||
| 1007 | }, |
||
| 1008 | { |
||
| 1009 | "cell_type": "code", |
||
| 1010 | "execution_count": null, |
||
| 1011 | "metadata": {}, |
||
| 1012 | "outputs": [], |
||
| 1013 | "source": [ |
||
| 1014 | "train_df['RECORD_TIME_DAY'] = train_df['RECORD_TIME'].map(lambda x: datetime.datetime.strptime(x, '%Y-%m-%d %H:%M:%S').strftime('%Y-%m-%d'))\n", |
||
| 1015 | "train_df['RECORD_TIME_HOUR'] = train_df['RECORD_TIME'].map(lambda x: datetime.datetime.strptime(x, '%Y-%m-%d %H:%M:%S').strftime('%Y-%m-%d %H'))\n", |
||
| 1016 | "train_df.head()" |
||
| 1017 | ] |
||
| 1018 | }, |
||
| 1019 | { |
||
| 1020 | "cell_type": "code", |
||
| 1021 | "execution_count": null, |
||
| 1022 | "metadata": {}, |
||
| 1023 | "outputs": [], |
||
| 1024 | "source": [ |
||
| 1025 | "train_df_day = train_df.groupby(['PATIENT_ID', 'ADMISSION_DATE', 'DEPARTURE_DATE', 'RECORD_TIME_DAY'], dropna=True, as_index = False).mean()\n", |
||
| 1026 | "train_df_hour = train_df.groupby(['PATIENT_ID', 'ADMISSION_DATE', 'DEPARTURE_DATE', 'RECORD_TIME_HOUR'], dropna=True, as_index = False).mean()\n", |
||
| 1027 | "\n", |
||
| 1028 | "len(train_df), len(train_df_day), len(train_df_hour)" |
||
| 1029 | ] |
||
| 1030 | }, |
||
| 1031 | { |
||
| 1032 | "attachments": {}, |
||
| 1033 | "cell_type": "markdown", |
||
| 1034 | "metadata": {}, |
||
| 1035 | "source": [ |
||
| 1036 | "\n", |
||
| 1037 | "```\n", |
||
| 1038 | "number of visits (total)\n", |
||
| 1039 | "- Original data: 168777\n", |
||
| 1040 | "- Merge by hour: 130141\n", |
||
| 1041 | "- Merge by day: 42204\n", |
||
| 1042 | "```" |
||
| 1043 | ] |
||
| 1044 | }, |
||
| 1045 | { |
||
| 1046 | "cell_type": "code", |
||
| 1047 | "execution_count": null, |
||
| 1048 | "metadata": {}, |
||
| 1049 | "outputs": [], |
||
| 1050 | "source": [ |
||
| 1051 | "len(train_df['PATIENT_ID'].unique())" |
||
| 1052 | ] |
||
| 1053 | }, |
||
| 1054 | { |
||
| 1055 | "cell_type": "code", |
||
| 1056 | "execution_count": null, |
||
| 1057 | "metadata": {}, |
||
| 1058 | "outputs": [], |
||
| 1059 | "source": [ |
||
| 1060 | "def get_visit_intervals(df):\n", |
||
| 1061 | " ls = []\n", |
||
| 1062 | " for pat in df['PATIENT_ID'].unique():\n", |
||
| 1063 | " ls.append(len(df[df['PATIENT_ID'] == pat]))\n", |
||
| 1064 | " return ls" |
||
| 1065 | ] |
||
| 1066 | }, |
||
| 1067 | { |
||
| 1068 | "cell_type": "code", |
||
| 1069 | "execution_count": null, |
||
| 1070 | "metadata": {}, |
||
| 1071 | "outputs": [], |
||
| 1072 | "source": [ |
||
| 1073 | "ls_org = get_visit_intervals(train_df)\n", |
||
| 1074 | "ls_hour = get_visit_intervals(train_df_hour)\n", |
||
| 1075 | "ls_day = get_visit_intervals(train_df_day)" |
||
| 1076 | ] |
||
| 1077 | }, |
||
| 1078 | { |
||
| 1079 | "cell_type": "code", |
||
| 1080 | "execution_count": null, |
||
| 1081 | "metadata": {}, |
||
| 1082 | "outputs": [], |
||
| 1083 | "source": [ |
||
| 1084 | "import matplotlib.pyplot as plt\n", |
||
| 1085 | "from matplotlib.ticker import PercentFormatter\n", |
||
| 1086 | "import matplotlib.font_manager as font_manager\n", |
||
| 1087 | "import pandas as pd\n", |
||
| 1088 | "import numpy as np\n", |
||
| 1089 | "csfont = {'fontname':'Times New Roman', 'fontsize': 18}\n", |
||
| 1090 | "font = 'Times New Roman'\n", |
||
| 1091 | "fig=plt.figure(figsize=(18,4), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 1092 | "plt.style.use('seaborn-whitegrid')\n", |
||
| 1093 | "color = 'cornflowerblue'\n", |
||
| 1094 | "ec = 'None'\n", |
||
| 1095 | "alpha=0.5\n", |
||
| 1096 | "\n", |
||
| 1097 | "ax = plt.subplot(1, 3, 1)\n", |
||
| 1098 | "ax.hist(ls_org, bins=20, weights=np.ones(len(ls_org)) / len(ls_org), color=color, ec=ec, alpha=alpha, label='overall')\n", |
||
| 1099 | "plt.xlabel('Num of visits (org)',**csfont)\n", |
||
| 1100 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1101 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1102 | "plt.xticks(**csfont)\n", |
||
| 1103 | "plt.yticks(**csfont)\n", |
||
| 1104 | "\n", |
||
| 1105 | "ax = plt.subplot(1, 3, 2)\n", |
||
| 1106 | "ax.hist(ls_hour, bins=20, weights=np.ones(len(ls_hour)) / len(ls_hour), color=color, ec=ec, alpha=alpha, label='overall')\n", |
||
| 1107 | "plt.xlabel('Num of visits (hour)',**csfont)\n", |
||
| 1108 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1109 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1110 | "plt.xticks(**csfont)\n", |
||
| 1111 | "plt.yticks(**csfont)\n", |
||
| 1112 | "\n", |
||
| 1113 | "ax = plt.subplot(1, 3, 3)\n", |
||
| 1114 | "ax.hist(ls_day, bins=20, weights=np.ones(len(ls_day)) / len(ls_day), color=color, ec=ec, alpha=alpha, label='overall')\n", |
||
| 1115 | "plt.xlabel('Num of visits (day)',**csfont)\n", |
||
| 1116 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1117 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1118 | "plt.xticks(**csfont)\n", |
||
| 1119 | "plt.yticks(**csfont)\n", |
||
| 1120 | "\n", |
||
| 1121 | "plt.show()" |
||
| 1122 | ] |
||
| 1123 | }, |
||
| 1124 | { |
||
| 1125 | "cell_type": "code", |
||
| 1126 | "execution_count": null, |
||
| 1127 | "metadata": {}, |
||
| 1128 | "outputs": [], |
||
| 1129 | "source": [ |
||
| 1130 | "def get_statistic(lst, name):\n", |
||
| 1131 | " print(f'[{name}]\\tMax:\\t{max(lst)}, Min:\\t{min(lst)}, Median:\\t{np.median(lst)}, Mean:\\t{np.mean(lst)}, 80%:\\t{np.quantile(lst, 0.8)}, 90%:\\t{np.quantile(lst, 0.9)}, 95%:\\t{np.quantile(lst, 0.95)}')" |
||
| 1132 | ] |
||
| 1133 | }, |
||
| 1134 | { |
||
| 1135 | "cell_type": "code", |
||
| 1136 | "execution_count": null, |
||
| 1137 | "metadata": {}, |
||
| 1138 | "outputs": [], |
||
| 1139 | "source": [ |
||
| 1140 | "get_statistic(ls_org, 'ls_org')\n", |
||
| 1141 | "get_statistic(ls_hour, 'ls_hour')\n", |
||
| 1142 | "get_statistic(ls_day, 'ls_day')" |
||
| 1143 | ] |
||
| 1144 | }, |
||
| 1145 | { |
||
| 1146 | "cell_type": "code", |
||
| 1147 | "execution_count": null, |
||
| 1148 | "metadata": {}, |
||
| 1149 | "outputs": [], |
||
| 1150 | "source": [ |
||
| 1151 | "train_df_hour['LOS'] = train_df_hour['ADMISSION_DATE']\n", |
||
| 1152 | "train_df_hour['LOS_HOUR'] = train_df_hour['ADMISSION_DATE']" |
||
| 1153 | ] |
||
| 1154 | }, |
||
| 1155 | { |
||
| 1156 | "cell_type": "code", |
||
| 1157 | "execution_count": null, |
||
| 1158 | "metadata": {}, |
||
| 1159 | "outputs": [], |
||
| 1160 | "source": [ |
||
| 1161 | "train_df_hour = train_df_hour.reset_index()" |
||
| 1162 | ] |
||
| 1163 | }, |
||
| 1164 | { |
||
| 1165 | "cell_type": "code", |
||
| 1166 | "execution_count": null, |
||
| 1167 | "metadata": {}, |
||
| 1168 | "outputs": [], |
||
| 1169 | "source": [ |
||
| 1170 | "for idx in tqdm(range(len(train_df_hour))):\n", |
||
| 1171 | " info = train_df_hour.loc[idx]\n", |
||
| 1172 | " admission = datetime.datetime.strptime(info['ADMISSION_DATE'], '%Y-%m-%d %H:%M:%S')\n", |
||
| 1173 | " departure = datetime.datetime.strptime(info['DEPARTURE_DATE'], '%Y-%m-%d %H:%M:%S')\n", |
||
| 1174 | " visit_hour = datetime.datetime.strptime(info['RECORD_TIME_HOUR'], '%Y-%m-%d %H')\n", |
||
| 1175 | " hour = (departure - visit_hour).seconds / (24 * 60 * 60) + (departure - visit_hour).days\n", |
||
| 1176 | " los = (departure - admission).seconds / (24 * 60 * 60) + (departure - admission).days\n", |
||
| 1177 | " train_df_hour.at[idx, 'LOS'] = float(los)\n", |
||
| 1178 | " train_df_hour.at[idx, 'LOS_HOUR'] = float(hour)" |
||
| 1179 | ] |
||
| 1180 | }, |
||
| 1181 | { |
||
| 1182 | "cell_type": "code", |
||
| 1183 | "execution_count": null, |
||
| 1184 | "metadata": {}, |
||
| 1185 | "outputs": [], |
||
| 1186 | "source": [ |
||
| 1187 | "train_df_hour['LOS']" |
||
| 1188 | ] |
||
| 1189 | }, |
||
| 1190 | { |
||
| 1191 | "cell_type": "code", |
||
| 1192 | "execution_count": null, |
||
| 1193 | "metadata": {}, |
||
| 1194 | "outputs": [], |
||
| 1195 | "source": [ |
||
| 1196 | "los = []\n", |
||
| 1197 | "for pat in tqdm(train_df_hour['PATIENT_ID'].unique()):\n", |
||
| 1198 | " los.append(float(train_df_hour[train_df_hour['PATIENT_ID'] == pat]['LOS'].head(1)))" |
||
| 1199 | ] |
||
| 1200 | }, |
||
| 1201 | { |
||
| 1202 | "cell_type": "code", |
||
| 1203 | "execution_count": null, |
||
| 1204 | "metadata": {}, |
||
| 1205 | "outputs": [], |
||
| 1206 | "source": [ |
||
| 1207 | "get_statistic(los, 'los')" |
||
| 1208 | ] |
||
| 1209 | }, |
||
| 1210 | { |
||
| 1211 | "cell_type": "code", |
||
| 1212 | "execution_count": null, |
||
| 1213 | "metadata": {}, |
||
| 1214 | "outputs": [], |
||
| 1215 | "source": [ |
||
| 1216 | "import matplotlib.pyplot as plt\n", |
||
| 1217 | "from matplotlib.ticker import PercentFormatter\n", |
||
| 1218 | "import matplotlib.font_manager as font_manager\n", |
||
| 1219 | "import pandas as pd\n", |
||
| 1220 | "import numpy as np\n", |
||
| 1221 | "csfont = {'fontname':'Times New Roman', 'fontsize': 18}\n", |
||
| 1222 | "font = 'Times New Roman'\n", |
||
| 1223 | "fig=plt.figure(figsize=(6,6), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 1224 | "plt.style.use('seaborn-whitegrid')\n", |
||
| 1225 | "color = 'cornflowerblue'\n", |
||
| 1226 | "ec = 'None'\n", |
||
| 1227 | "alpha=0.5\n", |
||
| 1228 | "\n", |
||
| 1229 | "ax = plt.subplot(1, 1, 1)\n", |
||
| 1230 | "ax.hist(los, bins=20, weights=np.ones(len(los)) / len(los), color=color, ec=ec, alpha=alpha, label='overall')\n", |
||
| 1231 | "plt.xlabel('Length of stay',**csfont)\n", |
||
| 1232 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1233 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1234 | "plt.xticks(**csfont)\n", |
||
| 1235 | "plt.yticks(**csfont)\n", |
||
| 1236 | "\n", |
||
| 1237 | "plt.show()" |
||
| 1238 | ] |
||
| 1239 | }, |
||
| 1240 | { |
||
| 1241 | "cell_type": "code", |
||
| 1242 | "execution_count": null, |
||
| 1243 | "metadata": {}, |
||
| 1244 | "outputs": [], |
||
| 1245 | "source": [ |
||
| 1246 | "train_df_hour_idx = train_df_hour.reset_index()" |
||
| 1247 | ] |
||
| 1248 | }, |
||
| 1249 | { |
||
| 1250 | "cell_type": "code", |
||
| 1251 | "execution_count": null, |
||
| 1252 | "metadata": {}, |
||
| 1253 | "outputs": [], |
||
| 1254 | "source": [ |
||
| 1255 | "train_df_hour_idx['LOS'] = train_df_hour_idx['ADMISSION_DATE']\n", |
||
| 1256 | "\n", |
||
| 1257 | "for idx in tqdm(range(len(train_df_hour_idx))):\n", |
||
| 1258 | " info = train_df_hour_idx.loc[idx]\n", |
||
| 1259 | " # admission = datetime.datetime.strptime(info['ADMISSION_DATE'], '%Y-%m-%d %H:%M:%S')\n", |
||
| 1260 | " departure = datetime.datetime.strptime(info['DEPARTURE_DATE'], '%Y-%m-%d %H:%M:%S')\n", |
||
| 1261 | " visit_hour = datetime.datetime.strptime(info['RECORD_TIME_HOUR'], '%Y-%m-%d %H')\n", |
||
| 1262 | " hour = (departure - visit_hour).seconds / (24 * 60 * 60) + (departure - visit_hour).days\n", |
||
| 1263 | " train_df_hour_idx.at[idx, 'LOS'] = float(hour)" |
||
| 1264 | ] |
||
| 1265 | }, |
||
| 1266 | { |
||
| 1267 | "cell_type": "code", |
||
| 1268 | "execution_count": null, |
||
| 1269 | "metadata": {}, |
||
| 1270 | "outputs": [], |
||
| 1271 | "source": [ |
||
| 1272 | "train_df_hour['LOS'] = train_df_hour['LOS_HOUR']\n", |
||
| 1273 | "train_df_hour.drop(columns=['LOS_HOUR'])" |
||
| 1274 | ] |
||
| 1275 | }, |
||
| 1276 | { |
||
| 1277 | "cell_type": "code", |
||
| 1278 | "execution_count": null, |
||
| 1279 | "metadata": {}, |
||
| 1280 | "outputs": [], |
||
| 1281 | "source": [ |
||
| 1282 | "# los_threshold = 13.0\n", |
||
| 1283 | "\n", |
||
| 1284 | "# visit_num_hour = []\n", |
||
| 1285 | "\n", |
||
| 1286 | "# for pat in tqdm(train_df_hour_idx['PATIENT_ID'].unique()):\n", |
||
| 1287 | "# pat_records = train_df_hour_idx[train_df_hour_idx['PATIENT_ID'] == pat]\n", |
||
| 1288 | "# hour = 0\n", |
||
| 1289 | "# for vis in pat_records.index:\n", |
||
| 1290 | "# pat_visit = pat_records.loc[vis]\n", |
||
| 1291 | "# if pat_visit['LOS_HOUR'] <= los_threshold:\n", |
||
| 1292 | "# hour += 1\n", |
||
| 1293 | "# visit_num_hour.append(hour)\n", |
||
| 1294 | "# if hour == 0:\n", |
||
| 1295 | "# print(pat)" |
||
| 1296 | ] |
||
| 1297 | }, |
||
| 1298 | { |
||
| 1299 | "cell_type": "code", |
||
| 1300 | "execution_count": null, |
||
| 1301 | "metadata": {}, |
||
| 1302 | "outputs": [], |
||
| 1303 | "source": [ |
||
| 1304 | "# import matplotlib.pyplot as plt\n", |
||
| 1305 | "# from matplotlib.ticker import PercentFormatter\n", |
||
| 1306 | "# import matplotlib.font_manager as font_manager\n", |
||
| 1307 | "# import pandas as pd\n", |
||
| 1308 | "# import numpy as np\n", |
||
| 1309 | "# csfont = {'fontname':'Times New Roman', 'fontsize': 18}\n", |
||
| 1310 | "# font = 'Times New Roman'\n", |
||
| 1311 | "# fig=plt.figure(figsize=(6,6), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 1312 | "# plt.style.use('seaborn-whitegrid')\n", |
||
| 1313 | "# color = 'cornflowerblue'\n", |
||
| 1314 | "# ec = 'None'\n", |
||
| 1315 | "# alpha=0.5\n", |
||
| 1316 | "\n", |
||
| 1317 | "# ax = plt.subplot(1, 1, 1)\n", |
||
| 1318 | "# ax.hist(visit_num_hour, bins=20, weights=np.ones(len(visit_num_hour)) / len(visit_num_hour), color=color, ec=ec, alpha=alpha, label='overall')\n", |
||
| 1319 | "# plt.xlabel('Visit num (80% los)',**csfont)\n", |
||
| 1320 | "# plt.ylabel('Percentage',**csfont)\n", |
||
| 1321 | "# plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1322 | "# plt.xticks(**csfont)\n", |
||
| 1323 | "# plt.yticks(**csfont)\n", |
||
| 1324 | "\n", |
||
| 1325 | "# plt.show()" |
||
| 1326 | ] |
||
| 1327 | }, |
||
| 1328 | { |
||
| 1329 | "cell_type": "code", |
||
| 1330 | "execution_count": null, |
||
| 1331 | "metadata": {}, |
||
| 1332 | "outputs": [], |
||
| 1333 | "source": [ |
||
| 1334 | "train_df = train_df_hour\n", |
||
| 1335 | "train_df.head()" |
||
| 1336 | ] |
||
| 1337 | }, |
||
| 1338 | { |
||
| 1339 | "cell_type": "code", |
||
| 1340 | "execution_count": null, |
||
| 1341 | "metadata": {}, |
||
| 1342 | "outputs": [], |
||
| 1343 | "source": [ |
||
| 1344 | "train_df.describe()" |
||
| 1345 | ] |
||
| 1346 | }, |
||
| 1347 | { |
||
| 1348 | "cell_type": "code", |
||
| 1349 | "execution_count": null, |
||
| 1350 | "metadata": {}, |
||
| 1351 | "outputs": [], |
||
| 1352 | "source": [ |
||
| 1353 | "get_statistic(train_df['LOS'], 'los')" |
||
| 1354 | ] |
||
| 1355 | }, |
||
| 1356 | { |
||
| 1357 | "cell_type": "code", |
||
| 1358 | "execution_count": null, |
||
| 1359 | "metadata": {}, |
||
| 1360 | "outputs": [], |
||
| 1361 | "source": [ |
||
| 1362 | "train_df['LOS'] = train_df['LOS'].clip(lower=0)" |
||
| 1363 | ] |
||
| 1364 | }, |
||
| 1365 | { |
||
| 1366 | "cell_type": "code", |
||
| 1367 | "execution_count": null, |
||
| 1368 | "metadata": {}, |
||
| 1369 | "outputs": [], |
||
| 1370 | "source": [ |
||
| 1371 | "get_statistic(train_df['LOS'], 'los')" |
||
| 1372 | ] |
||
| 1373 | }, |
||
| 1374 | { |
||
| 1375 | "cell_type": "code", |
||
| 1376 | "execution_count": null, |
||
| 1377 | "metadata": {}, |
||
| 1378 | "outputs": [], |
||
| 1379 | "source": [ |
||
| 1380 | "# the first visit of each person\n", |
||
| 1381 | "def init_prev(prev):\n", |
||
| 1382 | " miss = []\n", |
||
| 1383 | " l = len(prev)\n", |
||
| 1384 | " for idx in range(l):\n", |
||
| 1385 | " #print(prev[idx])\n", |
||
| 1386 | " #print(type(prev[idx]))\n", |
||
| 1387 | " if np.isnan(prev[idx]): # there is no previous record\n", |
||
| 1388 | " prev[idx] = test_median[idx] # replace nan to median\n", |
||
| 1389 | " miss.append(1) # mark miss as 1\n", |
||
| 1390 | " else: # there is a previous record\n", |
||
| 1391 | " miss.append(0)\n", |
||
| 1392 | " return miss\n", |
||
| 1393 | "\n", |
||
| 1394 | "# the rest of the visits\n", |
||
| 1395 | "def fill_nan(cur, prev):\n", |
||
| 1396 | " l = len(prev)\n", |
||
| 1397 | " miss = []\n", |
||
| 1398 | " for idx in range(l):\n", |
||
| 1399 | " #print(cur[idx])\n", |
||
| 1400 | " if np.isnan(cur[idx]): # there is no record in current timestep\n", |
||
| 1401 | " cur[idx] = prev[idx] # cur <- prev\n", |
||
| 1402 | " miss.append(1)\n", |
||
| 1403 | " else: # there is a record in current timestep\n", |
||
| 1404 | " miss.append(0)\n", |
||
| 1405 | " return miss" |
||
| 1406 | ] |
||
| 1407 | }, |
||
| 1408 | { |
||
| 1409 | "cell_type": "code", |
||
| 1410 | "execution_count": null, |
||
| 1411 | "metadata": {}, |
||
| 1412 | "outputs": [], |
||
| 1413 | "source": [ |
||
| 1414 | "index = train_df.loc[0].index\n", |
||
| 1415 | "\n", |
||
| 1416 | "csv = dict()\n", |
||
| 1417 | "for key in ['PatientID', 'RecordTime', 'AdmissionTime', 'DischargeTime', 'Outcome', 'LOS', 'Sex', 'Age']:\n", |
||
| 1418 | " csv[key] = []\n", |
||
| 1419 | "for key in index[8:-2]:\n", |
||
| 1420 | " csv[key] = []\n", |
||
| 1421 | " \n", |
||
| 1422 | "for pat in tqdm(patient_ids): # for all patients\n", |
||
| 1423 | " # get visits for pat.id == PATIENT_ID\n", |
||
| 1424 | " info = train_df[train_df['PATIENT_ID'] == pat]\n", |
||
| 1425 | " info = info[max(0, len(info) - 76):]\n", |
||
| 1426 | " idxs = info.index\n", |
||
| 1427 | " for i in idxs:\n", |
||
| 1428 | " visit = info.loc[i]\n", |
||
| 1429 | " for key in index[8:-2]:\n", |
||
| 1430 | " csv[key].append(visit[key])\n", |
||
| 1431 | " # ['PatientID', 'RecordTime', 'AdmissionTime', 'DischargeTime', 'Outcome', 'LOS', 'Sex', 'Age']\n", |
||
| 1432 | " csv['PatientID'].append(visit['PATIENT_ID'])\n", |
||
| 1433 | " t, h = visit['RECORD_TIME_HOUR'].split()\n", |
||
| 1434 | " t = t.split('-')\n", |
||
| 1435 | " csv['RecordTime'].append(t[1]+'/'+t[2]+'/'+t[0]+' '+h) # 2020-04-06 10 -> 04/06/2020 10\n", |
||
| 1436 | " t = visit['ADMISSION_DATE'][:10].split('-')\n", |
||
| 1437 | " csv['AdmissionTime'].append(t[1]+'/'+t[2]+'/'+t[0])\n", |
||
| 1438 | " t = visit['DEPARTURE_DATE'][:10].split('-')\n", |
||
| 1439 | " csv['DischargeTime'].append(t[1]+'/'+t[2]+'/'+t[0])\n", |
||
| 1440 | " csv['Outcome'].append(visit['OUTCOME'])\n", |
||
| 1441 | " csv['LOS'].append(visit['LOS_HOUR'])\n", |
||
| 1442 | " csv['Sex'].append(visit['SEX'])\n", |
||
| 1443 | " csv['Age'].append(visit['AGE'])\n", |
||
| 1444 | " \n", |
||
| 1445 | "pd.DataFrame(csv).to_csv('processed_data/CDSL.csv')" |
||
| 1446 | ] |
||
| 1447 | }, |
||
| 1448 | { |
||
| 1449 | "cell_type": "code", |
||
| 1450 | "execution_count": null, |
||
| 1451 | "metadata": {}, |
||
| 1452 | "outputs": [], |
||
| 1453 | "source": [ |
||
| 1454 | "x, y, demo, x_lab_len, missing_mask = [], [], [], [], []\n", |
||
| 1455 | "\n", |
||
| 1456 | "for pat in tqdm(patient_ids): # for all patients\n", |
||
| 1457 | " # get visits for pat.id == PATIENT_ID\n", |
||
| 1458 | " info = train_df[train_df['PATIENT_ID'] == pat]\n", |
||
| 1459 | " info = info[max(0, len(info) - 76):]\n", |
||
| 1460 | " indexes = info.index\n", |
||
| 1461 | " visit = info.loc[indexes[0]] # get the first visit\n", |
||
| 1462 | "\n", |
||
| 1463 | " # demographic data\n", |
||
| 1464 | " demo.append([visit[k] for k in demo_cols])\n", |
||
| 1465 | " \n", |
||
| 1466 | " # label\n", |
||
| 1467 | " outcome = visit['OUTCOME']\n", |
||
| 1468 | " los = []\n", |
||
| 1469 | "\n", |
||
| 1470 | " # lab test & vital signs\n", |
||
| 1471 | " tests = []\n", |
||
| 1472 | " prev = visit[test_cols]\n", |
||
| 1473 | " miss = [] # missing matrix\n", |
||
| 1474 | " miss.append(init_prev(prev)) # fill nan for the first visit for every patient and add missing status to missing matrix\n", |
||
| 1475 | " # leave = datetime.datetime.strptime(visit['DEPARTURE_DATE'], '%Y-%m-%d %H:%M:%S')\n", |
||
| 1476 | " \n", |
||
| 1477 | " first = True\n", |
||
| 1478 | " for i in indexes:\n", |
||
| 1479 | " visit = info.loc[i]\n", |
||
| 1480 | " # now = datetime.datetime.strptime(visit['RECORD_TIME'], '%Y-%m-%d %H')\n", |
||
| 1481 | " cur = visit[test_cols]\n", |
||
| 1482 | " tmp = fill_nan(cur, prev) # fill nan for the rest of the visits\n", |
||
| 1483 | " if not first:\n", |
||
| 1484 | " miss.append(tmp) # add missing status to missing matrix\n", |
||
| 1485 | " tests.append(cur)\n", |
||
| 1486 | " # los_visit = (leave - now).days\n", |
||
| 1487 | " # if los_visit < 0:\n", |
||
| 1488 | " # los_visit = 0\n", |
||
| 1489 | " los.append(visit['LOS'])\n", |
||
| 1490 | " prev = cur\n", |
||
| 1491 | " first = False\n", |
||
| 1492 | "\n", |
||
| 1493 | " valid_visit = len(los)\n", |
||
| 1494 | " # outcome = [outcome] * valid_visit\n", |
||
| 1495 | " x_lab_len.append(valid_visit)\n", |
||
| 1496 | " missing_mask.append(miss) # append the patient's missing matrix to the total missing matrix\n", |
||
| 1497 | "\n", |
||
| 1498 | " # tests = np.pad(tests, ((0, max_visit - valid_visit), (0, 0)))\n", |
||
| 1499 | " # outcome = np.pad(outcome, (0, max_visit - valid_visit))\n", |
||
| 1500 | " # los = np.pad(los, (0, max_visit - valid_visit))\n", |
||
| 1501 | " \n", |
||
| 1502 | " y.append([outcome, los])\n", |
||
| 1503 | " x.append(tests)" |
||
| 1504 | ] |
||
| 1505 | }, |
||
| 1506 | { |
||
| 1507 | "cell_type": "code", |
||
| 1508 | "execution_count": null, |
||
| 1509 | "metadata": {}, |
||
| 1510 | "outputs": [], |
||
| 1511 | "source": [ |
||
| 1512 | "all_x = x\n", |
||
| 1513 | "all_x_demo = demo\n", |
||
| 1514 | "all_y = y\n", |
||
| 1515 | "all_missing_mask = missing_mask" |
||
| 1516 | ] |
||
| 1517 | }, |
||
| 1518 | { |
||
| 1519 | "cell_type": "code", |
||
| 1520 | "execution_count": null, |
||
| 1521 | "metadata": {}, |
||
| 1522 | "outputs": [], |
||
| 1523 | "source": [ |
||
| 1524 | "all_x_labtest = np.array(all_x, dtype=object)\n", |
||
| 1525 | "x_lab_length = [len(_) for _ in all_x_labtest]\n", |
||
| 1526 | "x_lab_length = torch.tensor(x_lab_length, dtype=torch.int)\n", |
||
| 1527 | "max_length = int(x_lab_length.max())\n", |
||
| 1528 | "all_x_labtest = [torch.tensor(_) for _ in all_x_labtest]\n", |
||
| 1529 | "all_x_labtest = torch.nn.utils.rnn.pad_sequence((all_x_labtest), batch_first=True)\n", |
||
| 1530 | "all_x_demographic = torch.tensor(all_x_demo)\n", |
||
| 1531 | "batch_size, demo_dim = all_x_demographic.shape\n", |
||
| 1532 | "all_x_demographic = torch.reshape(all_x_demographic.repeat(1, max_length), (batch_size, max_length, demo_dim))\n", |
||
| 1533 | "all_x = torch.cat((all_x_demographic, all_x_labtest), 2)\n", |
||
| 1534 | "\n", |
||
| 1535 | "all_y = np.array(all_y, dtype=object)\n", |
||
| 1536 | "patient_list = []\n", |
||
| 1537 | "for pat in all_y:\n", |
||
| 1538 | " visits = []\n", |
||
| 1539 | " for i in pat[1]:\n", |
||
| 1540 | " visits.append([pat[0], i])\n", |
||
| 1541 | " patient_list.append(visits)\n", |
||
| 1542 | "new_all_y = np.array(patient_list, dtype=object)\n", |
||
| 1543 | "output_all_y = [torch.Tensor(_) for _ in new_all_y]\n", |
||
| 1544 | "output_all_y = torch.nn.utils.rnn.pad_sequence((output_all_y), batch_first=True)" |
||
| 1545 | ] |
||
| 1546 | }, |
||
| 1547 | { |
||
| 1548 | "cell_type": "code", |
||
| 1549 | "execution_count": null, |
||
| 1550 | "metadata": {}, |
||
| 1551 | "outputs": [], |
||
| 1552 | "source": [ |
||
| 1553 | "all_missing_mask = np.array(all_missing_mask, dtype=object)\n", |
||
| 1554 | "all_missing_mask = [torch.tensor(_) for _ in all_missing_mask]\n", |
||
| 1555 | "all_missing_mask = torch.nn.utils.rnn.pad_sequence((all_missing_mask), batch_first=True)" |
||
| 1556 | ] |
||
| 1557 | }, |
||
| 1558 | { |
||
| 1559 | "cell_type": "code", |
||
| 1560 | "execution_count": null, |
||
| 1561 | "metadata": {}, |
||
| 1562 | "outputs": [], |
||
| 1563 | "source": [ |
||
| 1564 | "all_x.shape" |
||
| 1565 | ] |
||
| 1566 | }, |
||
| 1567 | { |
||
| 1568 | "cell_type": "code", |
||
| 1569 | "execution_count": null, |
||
| 1570 | "metadata": {}, |
||
| 1571 | "outputs": [], |
||
| 1572 | "source": [ |
||
| 1573 | "all_missing_mask.shape" |
||
| 1574 | ] |
||
| 1575 | }, |
||
| 1576 | { |
||
| 1577 | "cell_type": "code", |
||
| 1578 | "execution_count": null, |
||
| 1579 | "metadata": {}, |
||
| 1580 | "outputs": [], |
||
| 1581 | "source": [ |
||
| 1582 | "# save pickle format dataset (torch)\n", |
||
| 1583 | "pd.to_pickle(all_x,f'./processed_data/x.pkl' )\n", |
||
| 1584 | "pd.to_pickle(all_missing_mask,f'./processed_data/missing_mask.pkl' )\n", |
||
| 1585 | "pd.to_pickle(output_all_y,f'./processed_data/y.pkl' )\n", |
||
| 1586 | "pd.to_pickle(x_lab_length,f'./processed_data/visits_length.pkl' )" |
||
| 1587 | ] |
||
| 1588 | }, |
||
| 1589 | { |
||
| 1590 | "cell_type": "code", |
||
| 1591 | "execution_count": null, |
||
| 1592 | "metadata": {}, |
||
| 1593 | "outputs": [], |
||
| 1594 | "source": [ |
||
| 1595 | "# Calculate patients' outcome statistics (patients-wise)\n", |
||
| 1596 | "outcome_list = []\n", |
||
| 1597 | "y_outcome = output_all_y[:, :, 0]\n", |
||
| 1598 | "indices = torch.arange(len(x_lab_length), dtype=torch.int64)\n", |
||
| 1599 | "for i in indices:\n", |
||
| 1600 | " outcome_list.append(y_outcome[i][0].item())\n", |
||
| 1601 | "outcome_list = np.array(outcome_list)\n", |
||
| 1602 | "print(len(outcome_list))\n", |
||
| 1603 | "unique, count=np.unique(outcome_list,return_counts=True)\n", |
||
| 1604 | "data_count=dict(zip(unique,count))\n", |
||
| 1605 | "print(data_count)" |
||
| 1606 | ] |
||
| 1607 | }, |
||
| 1608 | { |
||
| 1609 | "cell_type": "code", |
||
| 1610 | "execution_count": null, |
||
| 1611 | "metadata": {}, |
||
| 1612 | "outputs": [], |
||
| 1613 | "source": [ |
||
| 1614 | "# Calculate patients' outcome statistics (records-wise)\n", |
||
| 1615 | "outcome_records_list = []\n", |
||
| 1616 | "y_outcome = output_all_y[:, :, 0]\n", |
||
| 1617 | "indices = torch.arange(len(x_lab_length), dtype=torch.int64)\n", |
||
| 1618 | "for i in indices:\n", |
||
| 1619 | " outcome_records_list.extend(y_outcome[i][0:x_lab_length[i]].tolist())\n", |
||
| 1620 | "outcome_records_list = np.array(outcome_records_list)\n", |
||
| 1621 | "print(len(outcome_records_list))\n", |
||
| 1622 | "unique, count=np.unique(outcome_records_list,return_counts=True)\n", |
||
| 1623 | "data_count=dict(zip(unique,count))\n", |
||
| 1624 | "print(data_count)" |
||
| 1625 | ] |
||
| 1626 | }, |
||
| 1627 | { |
||
| 1628 | "cell_type": "code", |
||
| 1629 | "execution_count": null, |
||
| 1630 | "metadata": {}, |
||
| 1631 | "outputs": [], |
||
| 1632 | "source": [ |
||
| 1633 | "# Calculate patients' mean los and 95% percentile los\n", |
||
| 1634 | "los_list = []\n", |
||
| 1635 | "y_los = output_all_y[:, :, 1]\n", |
||
| 1636 | "indices = torch.arange(len(x_lab_length), dtype=torch.int64)\n", |
||
| 1637 | "for i in indices:\n", |
||
| 1638 | " # los_list.extend(y_los[i][: x_lab_length[i].long()].tolist())\n", |
||
| 1639 | " los_list.append(y_los[i][0].item())\n", |
||
| 1640 | "los_list = np.array(los_list)\n", |
||
| 1641 | "print(los_list.mean() * 0.5)\n", |
||
| 1642 | "print(np.median(los_list) * 0.5)\n", |
||
| 1643 | "print(np.percentile(los_list, 95))\n", |
||
| 1644 | "\n", |
||
| 1645 | "print('median:', np.median(los_list))\n", |
||
| 1646 | "print('Q1:', np.percentile(los_list, 25))\n", |
||
| 1647 | "print('Q3:', np.percentile(los_list, 75))" |
||
| 1648 | ] |
||
| 1649 | }, |
||
| 1650 | { |
||
| 1651 | "cell_type": "code", |
||
| 1652 | "execution_count": null, |
||
| 1653 | "metadata": {}, |
||
| 1654 | "outputs": [], |
||
| 1655 | "source": [ |
||
| 1656 | "los_alive_list = np.array([los_list[i] for i in range(len(los_list)) if outcome_list[i] == 0])\n", |
||
| 1657 | "los_dead_list = np.array([los_list[i] for i in range(len(los_list)) if outcome_list[i] == 1])\n", |
||
| 1658 | "print(len(los_alive_list))\n", |
||
| 1659 | "print(len(los_dead_list))\n", |
||
| 1660 | "\n", |
||
| 1661 | "print('[Alive]')\n", |
||
| 1662 | "print('median:', np.median(los_alive_list))\n", |
||
| 1663 | "print('Q1:', np.percentile(los_alive_list, 25))\n", |
||
| 1664 | "print('Q3:', np.percentile(los_alive_list, 75))\n", |
||
| 1665 | "\n", |
||
| 1666 | "print('[Dead]')\n", |
||
| 1667 | "print('median:', np.median(los_dead_list))\n", |
||
| 1668 | "print('Q1:', np.percentile(los_dead_list, 25))\n", |
||
| 1669 | "print('Q3:', np.percentile(los_dead_list, 75))" |
||
| 1670 | ] |
||
| 1671 | }, |
||
| 1672 | { |
||
| 1673 | "cell_type": "code", |
||
| 1674 | "execution_count": null, |
||
| 1675 | "metadata": {}, |
||
| 1676 | "outputs": [], |
||
| 1677 | "source": [ |
||
| 1678 | "cdsl_los_statistics = {\n", |
||
| 1679 | " 'overall': los_list,\n", |
||
| 1680 | " 'alive': los_alive_list,\n", |
||
| 1681 | " 'dead': los_dead_list\n", |
||
| 1682 | "}\n", |
||
| 1683 | "pd.to_pickle(cdsl_los_statistics, 'cdsl_los_statistics.pkl')" |
||
| 1684 | ] |
||
| 1685 | }, |
||
| 1686 | { |
||
| 1687 | "cell_type": "code", |
||
| 1688 | "execution_count": null, |
||
| 1689 | "metadata": {}, |
||
| 1690 | "outputs": [], |
||
| 1691 | "source": [ |
||
| 1692 | "# calculate visits length Median [Q1, Q3]\n", |
||
| 1693 | "visits_list = np.array(x_lab_length)\n", |
||
| 1694 | "visits_alive_list = np.array([x_lab_length[i] for i in range(len(x_lab_length)) if outcome_list[i] == 0])\n", |
||
| 1695 | "visits_dead_list = np.array([x_lab_length[i] for i in range(len(x_lab_length)) if outcome_list[i] == 1])\n", |
||
| 1696 | "print(len(visits_alive_list))\n", |
||
| 1697 | "print(len(visits_dead_list))\n", |
||
| 1698 | "\n", |
||
| 1699 | "print('[Total]')\n", |
||
| 1700 | "print('median:', np.median(visits_list))\n", |
||
| 1701 | "print('Q1:', np.percentile(visits_list, 25))\n", |
||
| 1702 | "print('Q3:', np.percentile(visits_list, 75))\n", |
||
| 1703 | "\n", |
||
| 1704 | "print('[Alive]')\n", |
||
| 1705 | "print('median:', np.median(visits_alive_list))\n", |
||
| 1706 | "print('Q1:', np.percentile(visits_alive_list, 25))\n", |
||
| 1707 | "print('Q3:', np.percentile(visits_alive_list, 75))\n", |
||
| 1708 | "\n", |
||
| 1709 | "print('[Dead]')\n", |
||
| 1710 | "print('median:', np.median(visits_dead_list))\n", |
||
| 1711 | "print('Q1:', np.percentile(visits_dead_list, 25))\n", |
||
| 1712 | "print('Q3:', np.percentile(visits_dead_list, 75))" |
||
| 1713 | ] |
||
| 1714 | }, |
||
| 1715 | { |
||
| 1716 | "cell_type": "code", |
||
| 1717 | "execution_count": null, |
||
| 1718 | "metadata": {}, |
||
| 1719 | "outputs": [], |
||
| 1720 | "source": [ |
||
| 1721 | "def check_nan(x):\n", |
||
| 1722 | " if np.isnan(np.sum(x.cpu().numpy())):\n", |
||
| 1723 | " print(\"some values from input are nan\")\n", |
||
| 1724 | " else:\n", |
||
| 1725 | " print(\"no nan\")" |
||
| 1726 | ] |
||
| 1727 | }, |
||
| 1728 | { |
||
| 1729 | "cell_type": "code", |
||
| 1730 | "execution_count": null, |
||
| 1731 | "metadata": {}, |
||
| 1732 | "outputs": [], |
||
| 1733 | "source": [ |
||
| 1734 | "check_nan(all_x)" |
||
| 1735 | ] |
||
| 1736 | }, |
||
| 1737 | { |
||
| 1738 | "cell_type": "markdown", |
||
| 1739 | "metadata": {}, |
||
| 1740 | "source": [ |
||
| 1741 | "# Draw Charts" |
||
| 1742 | ] |
||
| 1743 | }, |
||
| 1744 | { |
||
| 1745 | "cell_type": "markdown", |
||
| 1746 | "metadata": {}, |
||
| 1747 | "source": [ |
||
| 1748 | "## Import packages" |
||
| 1749 | ] |
||
| 1750 | }, |
||
| 1751 | { |
||
| 1752 | "cell_type": "code", |
||
| 1753 | "execution_count": null, |
||
| 1754 | "metadata": {}, |
||
| 1755 | "outputs": [], |
||
| 1756 | "source": [ |
||
| 1757 | "import matplotlib.pyplot as plt\n", |
||
| 1758 | "from matplotlib.ticker import PercentFormatter\n", |
||
| 1759 | "import matplotlib.font_manager as font_manager\n", |
||
| 1760 | "import pandas as pd\n", |
||
| 1761 | "import numpy as np\n", |
||
| 1762 | "\n", |
||
| 1763 | "plt.style.use('seaborn-whitegrid')\n", |
||
| 1764 | "color = 'cornflowerblue'\n", |
||
| 1765 | "ec = 'None'\n", |
||
| 1766 | "alpha=0.5\n", |
||
| 1767 | "alive_color = 'olivedrab'\n", |
||
| 1768 | "dead_color = 'orchid'" |
||
| 1769 | ] |
||
| 1770 | }, |
||
| 1771 | { |
||
| 1772 | "cell_type": "markdown", |
||
| 1773 | "metadata": {}, |
||
| 1774 | "source": [ |
||
| 1775 | "## Read data" |
||
| 1776 | ] |
||
| 1777 | }, |
||
| 1778 | { |
||
| 1779 | "cell_type": "code", |
||
| 1780 | "execution_count": null, |
||
| 1781 | "metadata": {}, |
||
| 1782 | "outputs": [], |
||
| 1783 | "source": [ |
||
| 1784 | "demographic.head()" |
||
| 1785 | ] |
||
| 1786 | }, |
||
| 1787 | { |
||
| 1788 | "cell_type": "code", |
||
| 1789 | "execution_count": null, |
||
| 1790 | "metadata": {}, |
||
| 1791 | "outputs": [], |
||
| 1792 | "source": [ |
||
| 1793 | "train = pd.read_csv('./train.csv')\n", |
||
| 1794 | "train['PATIENT_ID']=train['PATIENT_ID'].astype(str)\n", |
||
| 1795 | "demographic['PATIENT_ID']=demographic['PATIENT_ID'].astype(str)\n", |
||
| 1796 | "pat = {\n", |
||
| 1797 | " 'PATIENT_ID': train['PATIENT_ID'].unique()\n", |
||
| 1798 | "}\n", |
||
| 1799 | "pat = pd.DataFrame(pat)\n", |
||
| 1800 | "demo = pd.merge(demographic, pat, on='PATIENT_ID', how='inner')\n", |
||
| 1801 | "\n", |
||
| 1802 | "demo_alive = demo.loc[demo['OUTCOME'] == 0]\n", |
||
| 1803 | "demo_dead = demo.loc[demo['OUTCOME'] == 1]\n", |
||
| 1804 | "demo_overall = demo" |
||
| 1805 | ] |
||
| 1806 | }, |
||
| 1807 | { |
||
| 1808 | "cell_type": "code", |
||
| 1809 | "execution_count": null, |
||
| 1810 | "metadata": {}, |
||
| 1811 | "outputs": [], |
||
| 1812 | "source": [ |
||
| 1813 | "demo.to_csv('demo_overall.csv', index=False)\n", |
||
| 1814 | "demo_alive.to_csv('demo_alive.csv', index=False)\n", |
||
| 1815 | "demo_dead.to_csv('demo_dead.csv', index=False)" |
||
| 1816 | ] |
||
| 1817 | }, |
||
| 1818 | { |
||
| 1819 | "cell_type": "code", |
||
| 1820 | "execution_count": null, |
||
| 1821 | "metadata": {}, |
||
| 1822 | "outputs": [], |
||
| 1823 | "source": [ |
||
| 1824 | "patient = pd.DataFrame({\"PATIENT_ID\": (demo_alive['PATIENT_ID'].unique())})\n", |
||
| 1825 | "lab_tests_alive = pd.merge(lab_tests, patient, how='inner', on='PATIENT_ID')\n", |
||
| 1826 | "print(len(lab_tests_alive['PATIENT_ID'].unique()))\n", |
||
| 1827 | "\n", |
||
| 1828 | "patient = pd.DataFrame({\"PATIENT_ID\": (demo_dead['PATIENT_ID'].unique())})\n", |
||
| 1829 | "lab_tests_dead = pd.merge(lab_tests, patient, how='inner', on='PATIENT_ID')\n", |
||
| 1830 | "print(len(lab_tests_dead['PATIENT_ID'].unique()))\n", |
||
| 1831 | "\n", |
||
| 1832 | "patient = pd.DataFrame({\"PATIENT_ID\": (demo_overall['PATIENT_ID'].unique())})\n", |
||
| 1833 | "lab_tests_overall = pd.merge(lab_tests, patient, how='inner', on='PATIENT_ID')\n", |
||
| 1834 | "print(len(lab_tests_overall['PATIENT_ID'].unique()))" |
||
| 1835 | ] |
||
| 1836 | }, |
||
| 1837 | { |
||
| 1838 | "cell_type": "code", |
||
| 1839 | "execution_count": null, |
||
| 1840 | "metadata": {}, |
||
| 1841 | "outputs": [], |
||
| 1842 | "source": [ |
||
| 1843 | "patient = pd.DataFrame({\"PATIENT_ID\": (demo_alive['PATIENT_ID'].unique())})\n", |
||
| 1844 | "vital_signs_alive = pd.merge(vital_signs, patient, how='inner', on='PATIENT_ID')\n", |
||
| 1845 | "len(vital_signs_alive['PATIENT_ID'].unique())" |
||
| 1846 | ] |
||
| 1847 | }, |
||
| 1848 | { |
||
| 1849 | "cell_type": "code", |
||
| 1850 | "execution_count": null, |
||
| 1851 | "metadata": {}, |
||
| 1852 | "outputs": [], |
||
| 1853 | "source": [ |
||
| 1854 | "patient = pd.DataFrame({\"PATIENT_ID\": (demo_dead['PATIENT_ID'].unique())})\n", |
||
| 1855 | "vital_signs_dead = pd.merge(vital_signs, patient, how='inner', on='PATIENT_ID')\n", |
||
| 1856 | "len(vital_signs_dead['PATIENT_ID'].unique())" |
||
| 1857 | ] |
||
| 1858 | }, |
||
| 1859 | { |
||
| 1860 | "cell_type": "code", |
||
| 1861 | "execution_count": null, |
||
| 1862 | "metadata": {}, |
||
| 1863 | "outputs": [], |
||
| 1864 | "source": [ |
||
| 1865 | "patient = pd.DataFrame({\"PATIENT_ID\": (demo_overall['PATIENT_ID'].unique())})\n", |
||
| 1866 | "vital_signs_overall = pd.merge(vital_signs, patient, how='inner', on='PATIENT_ID')\n", |
||
| 1867 | "len(vital_signs_overall['PATIENT_ID'].unique())" |
||
| 1868 | ] |
||
| 1869 | }, |
||
| 1870 | { |
||
| 1871 | "cell_type": "code", |
||
| 1872 | "execution_count": null, |
||
| 1873 | "metadata": {}, |
||
| 1874 | "outputs": [], |
||
| 1875 | "source": [ |
||
| 1876 | "limit = 0.05\n", |
||
| 1877 | "\n", |
||
| 1878 | "csfont = {'fontname':'Times New Roman', 'fontsize': 18}\n", |
||
| 1879 | "font = 'Times New Roman'\n", |
||
| 1880 | "fig=plt.figure(figsize=(16,12), dpi= 100, facecolor='w', edgecolor='k')\n", |
||
| 1881 | "\n", |
||
| 1882 | "idx = 1\n", |
||
| 1883 | "\n", |
||
| 1884 | "key = 'AGE'\n", |
||
| 1885 | "low = demo_overall[key].quantile(limit)\n", |
||
| 1886 | "high = demo_overall[key].quantile(1 - limit)\n", |
||
| 1887 | "demo_AGE_overall = demo_overall[demo_overall[key].between(low, high)]\n", |
||
| 1888 | "demo_AGE_dead = demo_dead[demo_dead[key].between(low, high)]\n", |
||
| 1889 | "demo_AGE_alive = demo_alive[demo_alive[key].between(low, high)]\n", |
||
| 1890 | "ax = plt.subplot(4, 4, idx)\n", |
||
| 1891 | "ax.hist(demo_AGE_overall[key], bins=20, weights=np.ones(len(demo_AGE_overall[key])) / len(demo_AGE_overall), color=color, ec=ec, alpha=alpha, label='overall')\n", |
||
| 1892 | "plt.xlabel('Age',**csfont)\n", |
||
| 1893 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1894 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1895 | "# ax.title('Age Histogram', **csfont)\n", |
||
| 1896 | "ax.hist(demo_AGE_alive[key], bins=20, weights=np.ones(len(demo_AGE_alive[key])) / len(demo_AGE_alive), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2, label='alive')\n", |
||
| 1897 | "ax.hist(demo_AGE_dead[key], bins=20, weights=np.ones(len(demo_AGE_dead[key])) / len(demo_AGE_dead), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2, label='dead')\n", |
||
| 1898 | "plt.xticks(**csfont)\n", |
||
| 1899 | "plt.yticks(**csfont)\n", |
||
| 1900 | "idx += 1\n", |
||
| 1901 | "\n", |
||
| 1902 | "key = 'TEMPERATURE'\n", |
||
| 1903 | "low = vital_signs_overall[key].quantile(limit)\n", |
||
| 1904 | "high = vital_signs_overall[key].quantile(1 - limit)\n", |
||
| 1905 | "vs_TEMPERATURE_overall = vital_signs_overall[vital_signs_overall[key].between(low, high)]\n", |
||
| 1906 | "vs_TEMPERATURE_dead = vital_signs_dead[vital_signs_dead[key].between(low, high)]\n", |
||
| 1907 | "vs_TEMPERATURE_alive = vital_signs_alive[vital_signs_alive[key].between(low, high)]\n", |
||
| 1908 | "plt.subplot(4, 4, idx)\n", |
||
| 1909 | "plt.hist(vs_TEMPERATURE_overall['TEMPERATURE'], bins=20, weights=np.ones(len(vs_TEMPERATURE_overall)) / len(vs_TEMPERATURE_overall), color=color, ec=ec, alpha=alpha)\n", |
||
| 1910 | "plt.xlabel('Temperature',**csfont)\n", |
||
| 1911 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1912 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1913 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 1914 | "plt.hist(vs_TEMPERATURE_alive['TEMPERATURE'], bins=20, weights=np.ones(len(vs_TEMPERATURE_alive)) / len(vs_TEMPERATURE_alive), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1915 | "plt.hist(vs_TEMPERATURE_dead['TEMPERATURE'], bins=20, weights=np.ones(len(vs_TEMPERATURE_dead)) / len(vs_TEMPERATURE_dead), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1916 | "plt.xticks(**csfont)\n", |
||
| 1917 | "plt.yticks(**csfont)\n", |
||
| 1918 | "idx += 1\n", |
||
| 1919 | "\n", |
||
| 1920 | "# plt.subplot(4, 4, 3)\n", |
||
| 1921 | "# plt.hist(lab_tests_overall['CREA -- CREATININA'], bins=20, density=True, color=color, ec=ec, alpha=alpha)\n", |
||
| 1922 | "# plt.xlabel('CREA -- CREATININA',**csfont)\n", |
||
| 1923 | "# plt.ylabel('Percentage',**csfont)\n", |
||
| 1924 | "# plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1925 | "# # plt.title('Temperature Histogram', **csfont)\n", |
||
| 1926 | "# plt.hist(lab_tests_alive['CREA -- CREATININA'], bins=20, density=True, color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1927 | "# plt.hist(lab_tests_dead['CREA -- CREATININA'], bins=20, density=True, color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1928 | "# plt.xticks(**csfont)\n", |
||
| 1929 | "# plt.yticks(**csfont)\n", |
||
| 1930 | "\n", |
||
| 1931 | "key = 'CREA -- CREATININA'\n", |
||
| 1932 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 1933 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 1934 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 1935 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 1936 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 1937 | "plt.subplot(4, 4, idx)\n", |
||
| 1938 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 1939 | "plt.xlabel('CREA -- CREATININA',**csfont)\n", |
||
| 1940 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1941 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1942 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 1943 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1944 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1945 | "plt.xticks(**csfont)\n", |
||
| 1946 | "plt.yticks(**csfont)\n", |
||
| 1947 | "idx += 1\n", |
||
| 1948 | "\n", |
||
| 1949 | "key = 'HEM -- Hemat¡es'\n", |
||
| 1950 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 1951 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 1952 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 1953 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 1954 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 1955 | "plt.subplot(4, 4, idx)\n", |
||
| 1956 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 1957 | "plt.xlabel('HEM -- Hemat¡es',**csfont)\n", |
||
| 1958 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1959 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1960 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 1961 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1962 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1963 | "plt.xticks(**csfont)\n", |
||
| 1964 | "plt.yticks(**csfont)\n", |
||
| 1965 | "idx += 1\n", |
||
| 1966 | "\n", |
||
| 1967 | "key = 'LEUC -- Leucocitos'\n", |
||
| 1968 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 1969 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 1970 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 1971 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 1972 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 1973 | "plt.subplot(4, 4, idx)\n", |
||
| 1974 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 1975 | "plt.xlabel('LEUC -- Leucocitos',**csfont)\n", |
||
| 1976 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1977 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1978 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 1979 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1980 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1981 | "plt.xticks(**csfont)\n", |
||
| 1982 | "plt.yticks(**csfont)\n", |
||
| 1983 | "idx += 1\n", |
||
| 1984 | "\n", |
||
| 1985 | "key = 'PLAQ -- Recuento de plaquetas'\n", |
||
| 1986 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 1987 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 1988 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 1989 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 1990 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 1991 | "plt.subplot(4, 4, idx)\n", |
||
| 1992 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 1993 | "plt.xlabel('PLAQ -- Recuento de plaquetas',**csfont)\n", |
||
| 1994 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 1995 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 1996 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 1997 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1998 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 1999 | "plt.xticks(**csfont)\n", |
||
| 2000 | "plt.yticks(**csfont)\n", |
||
| 2001 | "idx += 1\n", |
||
| 2002 | "\n", |
||
| 2003 | "key = 'CHCM -- Conc. Hemoglobina Corpuscular Media'\n", |
||
| 2004 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2005 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2006 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2007 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2008 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2009 | "plt.subplot(4, 4, idx)\n", |
||
| 2010 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2011 | "plt.xlabel('CHCM',**csfont)\n", |
||
| 2012 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2013 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2014 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2015 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2016 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2017 | "plt.xticks(**csfont)\n", |
||
| 2018 | "plt.yticks(**csfont)\n", |
||
| 2019 | "idx += 1\n", |
||
| 2020 | "\n", |
||
| 2021 | "key = 'HCTO -- Hematocrito'\n", |
||
| 2022 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2023 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2024 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2025 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2026 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2027 | "plt.subplot(4, 4, idx)\n", |
||
| 2028 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2029 | "plt.xlabel('HCTO -- Hematocrito',**csfont)\n", |
||
| 2030 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2031 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2032 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2033 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2034 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2035 | "plt.xticks(**csfont)\n", |
||
| 2036 | "plt.yticks(**csfont)\n", |
||
| 2037 | "idx += 1\n", |
||
| 2038 | "\n", |
||
| 2039 | "key = 'VCM -- Volumen Corpuscular Medio'\n", |
||
| 2040 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2041 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2042 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2043 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2044 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2045 | "plt.subplot(4, 4, idx)\n", |
||
| 2046 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2047 | "plt.xlabel('VCM -- Volumen Corpuscular Medio',**csfont)\n", |
||
| 2048 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2049 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2050 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2051 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2052 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2053 | "plt.xticks(**csfont)\n", |
||
| 2054 | "plt.yticks(**csfont)\n", |
||
| 2055 | "idx += 1\n", |
||
| 2056 | "\n", |
||
| 2057 | "key = 'HGB -- Hemoglobina'\n", |
||
| 2058 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2059 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2060 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2061 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2062 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2063 | "plt.subplot(4, 4, idx)\n", |
||
| 2064 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2065 | "plt.xlabel('HGB -- Hemoglobina',**csfont)\n", |
||
| 2066 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2067 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2068 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2069 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2070 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2071 | "plt.xticks(**csfont)\n", |
||
| 2072 | "plt.yticks(**csfont)\n", |
||
| 2073 | "idx += 1\n", |
||
| 2074 | "\n", |
||
| 2075 | "key = 'HCM -- Hemoglobina Corpuscular Media'\n", |
||
| 2076 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2077 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2078 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2079 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2080 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2081 | "plt.subplot(4, 4, idx)\n", |
||
| 2082 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2083 | "plt.xlabel('HCM -- Hemoglobina Corpuscular Media',**csfont)\n", |
||
| 2084 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2085 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2086 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2087 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2088 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2089 | "plt.xticks(**csfont)\n", |
||
| 2090 | "plt.yticks(**csfont)\n", |
||
| 2091 | "idx += 1\n", |
||
| 2092 | "\n", |
||
| 2093 | "key = 'NEU -- Neutr¢filos'\n", |
||
| 2094 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2095 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2096 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2097 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2098 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2099 | "plt.subplot(4, 4, idx)\n", |
||
| 2100 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2101 | "plt.xlabel('NEU -- Neutr¢filos',**csfont)\n", |
||
| 2102 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2103 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2104 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2105 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2106 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2107 | "plt.xticks(**csfont)\n", |
||
| 2108 | "plt.yticks(**csfont)\n", |
||
| 2109 | "idx += 1\n", |
||
| 2110 | "\n", |
||
| 2111 | "key = 'NEU% -- Neutr¢filos %'\n", |
||
| 2112 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2113 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2114 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2115 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2116 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2117 | "plt.subplot(4, 4, idx)\n", |
||
| 2118 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2119 | "plt.xlabel('NEU% -- Neutr¢filos%',**csfont)\n", |
||
| 2120 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2121 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2122 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2123 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2124 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2125 | "plt.xticks(**csfont)\n", |
||
| 2126 | "plt.yticks(**csfont)\n", |
||
| 2127 | "idx += 1\n", |
||
| 2128 | "\n", |
||
| 2129 | "key = 'LIN -- Linfocitos'\n", |
||
| 2130 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2131 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2132 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2133 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2134 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2135 | "plt.subplot(4, 4, idx)\n", |
||
| 2136 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2137 | "plt.xlabel('LIN -- Linfocitos',**csfont)\n", |
||
| 2138 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2139 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2140 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2141 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2142 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2143 | "plt.xticks(**csfont)\n", |
||
| 2144 | "plt.yticks(**csfont)\n", |
||
| 2145 | "idx += 1\n", |
||
| 2146 | "\n", |
||
| 2147 | "key = 'LIN% -- Linfocitos %'\n", |
||
| 2148 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2149 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2150 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2151 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2152 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2153 | "plt.subplot(4, 4, idx)\n", |
||
| 2154 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2155 | "plt.xlabel('LIN% -- Linfocitos%',**csfont)\n", |
||
| 2156 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2157 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2158 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2159 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2160 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2161 | "plt.xticks(**csfont)\n", |
||
| 2162 | "plt.yticks(**csfont)\n", |
||
| 2163 | "idx += 1\n", |
||
| 2164 | "\n", |
||
| 2165 | "key = 'ADW -- Coeficiente de anisocitosis'\n", |
||
| 2166 | "low = lab_tests_overall[key].quantile(limit)\n", |
||
| 2167 | "high = lab_tests_overall[key].quantile(1 - limit)\n", |
||
| 2168 | "lt_key_overall = lab_tests_overall[lab_tests_overall[key].between(low, high)]\n", |
||
| 2169 | "lt_key_dead = lab_tests_dead[lab_tests_dead[key].between(low, high)]\n", |
||
| 2170 | "lt_key_alive = lab_tests_alive[lab_tests_alive[key].between(low, high)]\n", |
||
| 2171 | "plt.subplot(4, 4, idx)\n", |
||
| 2172 | "plt.hist(lt_key_overall[key], bins=20, weights=np.ones(len(lt_key_overall[key])) / len(lt_key_overall[key]), color=color, ec=ec, alpha=alpha)\n", |
||
| 2173 | "plt.xlabel('ADW -- Coeficiente de anisocitosis',**csfont)\n", |
||
| 2174 | "plt.ylabel('Percentage',**csfont)\n", |
||
| 2175 | "plt.gca().yaxis.set_major_formatter(PercentFormatter(1))\n", |
||
| 2176 | "# plt.title('Temperature Histogram', **csfont)\n", |
||
| 2177 | "plt.hist(lt_key_alive[key], bins=20, weights=np.ones(len(lt_key_alive[key])) / len(lt_key_alive[key]), color='green', ec=alive_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2178 | "plt.hist(lt_key_dead[key], bins=20, weights=np.ones(len(lt_key_dead[key])) / len(lt_key_dead[key]), color='green', ec=dead_color, alpha=1, histtype=\"step\", linewidth=2)\n", |
||
| 2179 | "plt.xticks(**csfont)\n", |
||
| 2180 | "plt.yticks(**csfont)\n", |
||
| 2181 | "idx += 1\n", |
||
| 2182 | "\n", |
||
| 2183 | "handles, labels = ax.get_legend_handles_labels()\n", |
||
| 2184 | "print(handles, labels)\n", |
||
| 2185 | "# fig.legend(handles, labels, loc='upper center')\n", |
||
| 2186 | "plt.figlegend(handles, labels, loc='upper center', ncol=5, fontsize=18, bbox_to_anchor=(0.5, 1.05), prop=font_manager.FontProperties(family='Times New Roman',\n", |
||
| 2187 | " style='normal', size=18))\n", |
||
| 2188 | "# fig.legend(, [], loc='upper center')\n", |
||
| 2189 | "\n", |
||
| 2190 | "fig.tight_layout()\n", |
||
| 2191 | "plt.show()" |
||
| 2192 | ] |
||
| 2193 | } |
||
| 2194 | ], |
||
| 2195 | "metadata": { |
||
| 2196 | "kernelspec": { |
||
| 2197 | "display_name": "Python 3.7.11 ('python37')", |
||
| 2198 | "language": "python", |
||
| 2199 | "name": "python3" |
||
| 2200 | }, |
||
| 2201 | "language_info": { |
||
| 2202 | "codemirror_mode": { |
||
| 2203 | "name": "ipython", |
||
| 2204 | "version": 3 |
||
| 2205 | }, |
||
| 2206 | "file_extension": ".py", |
||
| 2207 | "mimetype": "text/x-python", |
||
| 2208 | "name": "python", |
||
| 2209 | "nbconvert_exporter": "python", |
||
| 2210 | "pygments_lexer": "ipython3", |
||
| 2211 | "version": "3.7.11" |
||
| 2212 | }, |
||
| 2213 | "vscode": { |
||
| 2214 | "interpreter": { |
||
| 2215 | "hash": "a10b846bdc9fc41ee38835cbc29d70b69dd5fd54e1341ea2c410a7804a50447a" |
||
| 2216 | } |
||
| 2217 | } |
||
| 2218 | }, |
||
| 2219 | "nbformat": 4, |
||
| 2220 | "nbformat_minor": 2 |
||
| 2221 | } |