Banks get many customers looking for loans everyday. This gives them access to information that can be leveraged to create machine learning solutions.
- Make more better and more accurate loan decisions
- Find connections, correlations, and biases in their decision-making process
- Stream Line and standardize Application Process
- Reduce customer support overhead and improve customer communication
- Data Exploration Notebook:
- Use matplotlib and seaborn to visualize data, find connections and a little story telling
- Modeling Notebook:
- Create a Pipeline to preprocess the data
- Test multiple classification algorithms
- Saving the best model and preprocessor with pickle library
- Deployment:
- Flask request app.
- simple script run through the terminal and can be accessed through the modeling notebook
- uses the pickled files to make a prediciton and return the prediction
- Flask GUI app.
- Script run through the terminal, but is accessed through a web browser
- Renders a GUI interface with the pages:
- Home page:
- takes inputs for variables present in the data
- button for submitting the information
- Result page:
- takes information submitted in the home page and uses the pickled model to make a prediction
- returns the prediction
- Home page:
- Flask request app.