Samuel-Kasusya/kenya-disease-dashboard

# Kenya Disease Burden — Interactive Dashboard **Student:** Samuel Mwanzia · **Course:** STA3050 A single-page Streamlit dashboard for a time-series term paper on Kenya's epidemiological transition (1980–2023): the shift between infectious and chronic (non-communicable) diseases, the HIV-driven **double crossover**, the **Lake Victoria** counties that never recovered, and **model accuracy**. ## Sections 1. **Context** — the epidemiological-transition question, national infectious-vs-chronic trend, GapMinder life-expectancy & child-mortality context. 2. **The Double Crossover** — HIV surge (1990) → NCD recovery (2010) in 42 of 47 counties, with a per-county explorer and the spatial transition map. 3. **The Lake Victoria Forecast** — the four HIV-hotspot counties still infectious-dominant in 2023, and ARIMA vs Prophet forecasts to 2043. 4. **Model Accuracy** — live stationarity tests (ADF/KPSS) and a live ARIMA hold-out backtest reporting MAE / RMSE / MAPE. ## Run locally ```bash pip install -r requirements.txt streamlit run app.py ``` Then open the URL Streamlit prints (usually localhost). ## Deploy a public link (Streamlit Community Cloud — free) 1. Sign in at share.streamlit.io with the GitHub account that owns this repo. 2. Click **New app** → select repo `Samuel-Kasusya/kenya-disease-dashboard`, branch `main`, main file `app.py`. 3. Click **Deploy**. The resulting `https:// .streamlit.app` URL is the clickable link to share with the lecturer. ## Project layout ``` app.py # the dashboard requirements.txt # dependencies .streamlit/config.toml # light theme Data/ # source datasets (IHME, GapMinder, KNBS, GeoJSON) Outputs/ # pre-rendered forecast & spatial-map charts Notebook/ # the source analysis notebook ``` ## Data sources - **IHME** Global Burden of Disease — death rates per 100,000 by cause, county and year. - **GapMinder** — life expectancy and child mo …