Throughout this training, we will use a small sample data set. If you follow the instructions of environment setup, you will be able to find the sample data in /bootcamp/data folder.
There are two files, case.csv and control.csv respectively. We define patient who developed heart failure (HF) at some time point as case patient and who didn't develop HF as control patient.
Each line of the sample data file is a tuple with format (patient-id, event-id, timestamp, value) like
020E860BD31CAC69,DRUG36987254604,968,30.0
020E860BD31CAC69,DRUG64158080642,974,30.0
020E860BD31CAC69,DRUG00440128228,976,60.0
020E860BD31CAC69,DIAG486,907,1.0
020E860BD31CAC69,DIAG7863,907,1.0
020E860BD31CAC69,DIAGV5866,907,1.0
020E860BD31CAC69,DIAG3659,907,1.0
020E860BD31CAC69,DIAGRG199,907,1.0
020E860BD31CAC69,PAYMENT,907,15000.0
020E860BD31CAC69,heartfailure,956,1.0
patient-id is just an identifier which don't have any meaning but to distinguish different patients. For example, the portion of data we shwo about is all about patient with id 020E860BD31CAC69.event-id encodes what happened to a patient. For example, DRUG00440128228 means a certain drug, DIAG486 means been diagnoised with Pneumonia and PAYMENT means made a payment.timestamp shows when does the event happen. Here the timestamp is just an offset from an unspecified start point for simplicity of processing and for privacy of patients.value is associated value of event. For drug it means the dosage, for payment means amount and for diagnostic type event like DIAG486 value equals 1 just means the event happened. The event heartfailure is a little bit differnt, for control patient, you will find event heartfailure have value equals 0 and for case patient equals 1. The above sample data shows the patient 020E860BD31CAC69 was diagnoised with heart failure at timestamp 956.