Heart Failure Clinical Records

Heart Failure Clinical Records

Approach

In this assignment, I selected a dataset from UCI Machine Learning Repository. The research looked at patients who already had heart failure and investigated whether clinical characteristics could help predict survival/death during the follow-up period. The dataset contains 12 features and 299 patients. The variables contains such as age, diabetes, blood pressure, ejection fraction, serum creatinine, smoking, follow-up time, and death event. I chose this dataset because it includes a clear outcome variable, death event which indicates whether a patient died during the follow-up period and can be used for future analysis.

My plan is load data into R. I will then select a smaller set of meaningful variables that can provide useful information about the patients and their outcomes.

To make the data easier to understand, I will rename variables that are not immediately intuitive and convert binary variables coded as 0 and 1 into descriptive labels such as “Yes” and “No.” I will also include the death_event variable as the target variable. Finally, I will check the resulting data frame to make sure the selected variables have appropriate names, values, and data types.

Data Source:

https://archive.ics.uci.edu/dataset/519/heart+failure+clinical+records

Research Question

Among people who already have heart failure, can information about their health and clinical measurements help us understand or predict who is more likely to survive?