Machine Learning Process from the Idea to the App (Stroke Prediction)

[frankmollard]

Hello all,

I created a tutorial where I show how to develop an app that includes machine learning algorithms. In particular, it highlights the difference to more deterministic projects. I do this using the example of predicting brain strokes. The idea is to develop an app that gives patients the probability of having a stroke by entering their data. Starting with the idea up to the finished app, every step will be gone through.

Here is the Kaggle Notebook:
https://www.kaggle.com/code/frankmollard/machine-learning-process-idea-2-app

Here’s where to find the Streamlit app:
https://stroke-risk-assessment.streamlit.app/

I appreciate any feedback

Best,

Frank

[sai_krishna1]

You app looks cool but if you provide drop down for the input values like age,marriage status it would be great.

And I think your data might have some issues, occupation type is showing as “Children”.

Overall a great app need some tuning here and there.

[frankmollard]

Dear Sai,
thank you for your feedback.

The marriage status is indeed already dropdown. I chose the sliders because I want the user to have fun using the app. This makes experimentation easier in particular with mobile phone.

Occupation = children is correct regarding the data - maybe a bit confusing. I have changed it to „child“ to make clear that being a child is meant. See Kaggle Notebook, everything is described there.

Best regards