I explored the Heart‑Disease‑Prediction‑using‑KNN repository on my GitHub, a Python project (3 stars, 1 fork) created and updated in July 2025, hosting a Streamlit app at https://heart-disease-knn-predictor.streamlit.app/. The goal is to predict heart disease using a K‑Nearest Neighbors classifier on the heart.csv dataset.
Missing values count = 0 (verified with df.isnull().sum() confirming there were none).
Duplicate rows count = 0 (checked via df.duplicated().sum() and found no duplicates).
Zero‑value imputation applied to Cholesterol and RestingBP using the mean of non‑zero values.
- Performed histograms with KDE for Age, RestingBP, Cholesterol, MaxHR.
- Created count plots for Sex, ChestPainType, FastingBS vs HeartDisease.
- Generated box & violin plots for Cholesterol and Age across HeartDisease classes.
- Visualized correlation matrix with a heatmap using
df.corr(numeric_only=True).