An interactive Streamlit application that lets users
- Explore a health dataset (EDA)
- Train & compare classification models (Models)
- Predict a person’s smoker status with a Decision-Tree classifier (Prediction)
- Provide an intuitive dashboard where anyone can:
- Visualise and interrogate the biosignal data on their own.
- Re-train several machine-learning models and benchmark their performance.
- Enter custom feature values and instantly obtain a smoker/non-smoker prediction.
| Item | Details |
|---|---|
| Source | Kaggle – “Smoker Status Prediction Using Biosignals” https://www.kaggle.com/datasets/gauravduttakiit/smoker-status-prediction-using-biosignals |
| Files used | train_dataset.csv |
| Target | smoking_status (binary: smoker = 1, non-smoker = 0) |
| Main features | Age, Height, Weight, BMI, Blood Pressure (sys/dia), Heart-rate, Body-Temp, Oxygen-Sat, etc. |
No preprocessing required: the dataset is already clean, well-formatted, and free of missing values or obvious outliers, so we load it directly into the app.
If you just want to try the dashboard and don’t care about creating a permanent virtual-env:
# ① Install uv (one-time, 5-second download)
curl -LsSf https://astral.sh/uv/install.sh | sh # macOS/Linux
# on Windows: iwr -useb https://astral.sh/uv/install.ps1 | iex
# ② Grab the code any way you like (clone / ZIP) and cd inside
git clone https://github.com/<your-username>/smokerstatusprediction.git
cd smokerstatusprediction # or the folder you un-zipped
# ③ Run the code
uv run streamlit run app.py