mohit / smoking drinking prediction
Smoking Drinking Prediction
Reading habits from health data.
You get a routine health check and a page of numbers back. Blood pressure, cholesterol, liver and blood markers, weight, vision. Each one sits in its own row with a normal range beside it, and nobody tells you what the numbers say about how you live.
They say more than you'd expect. Habits like smoking and drinking leave traces in the body, and those traces show up in ordinary check-up data.
I wanted to see how much a routine health check reveals about a person's habits, so I built this.
The model takes a person's health indicators and lifestyle data and predicts their smoking and drinking risk. Pandas cleans a large health screening dataset, Seaborn and Matplotlib show how each metric relates to the habits, and a logistic regression model does the classifying. A web app wraps it, so you can enter values and read the result in your browser.
Think of it as a lab report with a second opinion on lifestyle.
What you can do
- Enter health indicators and vitals and get a predicted risk.
- Run the prediction through logistic regression, which keeps every result easy to explain.
- Explore how each metric relates to smoking and drinking through charts.
- Follow the notebook from raw data to a trained model.
- Use the web app in your browser with nothing to install.
Where it stops
It predicts risk from patterns in a dataset and doesn't read behaviour directly. Many health markers move for reasons unrelated to smoking or drinking, so a high reading can have other causes. It isn't a diagnosis or medical advice. Use it to explore patterns, not to judge a person.
Built with Python, Scikit-learn, Pandas, Matplotlib and Seaborn.