Churn Prediction

FastAPI

Have implemented FastAPI in this project but due to some unfinished EDA I couldn't record the FastAPI user interface. Hence I have recorded it using StreamLit. Still you can find the screenshot of the FastAPI output below.

main.py, model.py, and random_forest_classifie.pkl are part of FastAPI

Overview

Churn Prediction is a Streamlit-based application designed to facilitate data analysis and machine learning tasks. It provides a user-friendly interface for data preprocessing, feature extraction, and model training, along with explanations of Random Forest's robustness.

Features

1. Home Page: Introduction to the application, with contact details and a motivational quote on data analysis.
2. Data PreProcessing:
   • Upload CSV files.
   • View data, check for null values, and fill missing values.
   • Detect and remove outliers.
   • Download the cleaned dataset.
   • View descriptive statistics and heatmaps of correlations.
3. Regression and Prediction:
   • Train Random Forest Classifier or Regressor models.
   • Save the trained model.
   • Predict outcomes based on user-provided input values.
4. Feature Extraction:
   • Delete unwanted columns.
   • Analyze feature importance using Random Forest.
5. Explanation:
   • Provides a detailed explanation of Random Forest's robustness to overfitting, outliers, and categorical values.

Installation

To run this app locally, follow these steps:

1. Clone the repository:

   git clone <repository-url>
   cd <repository-directory>

2. Install the required packages:

   pip install -r requirements.txt

3. Run the Streamlit app:

   streamlit run app.py

4. Run the FastAPI

   uvicorn main:app --reload

   Replace app.py with the name of your main Streamlit script if different.

5. Run the FastAPI

   uvicorn main:app --reload

   

Requirements

The requirements.txt file includes the necessary Python packages:

• base64
• numpy
• seaborn
• matplotlib
• streamlit
• pandas
• scikit-learn
• joblib

Usage

1. Home Page: Provides an introduction and contact information.
2. Data PreProcessing:
   • Upload a CSV file to preprocess.
   • The app checks for and fills null values, removes outliers, and allows downloading the cleaned data.
3. Regression and Prediction:
   • Choose between a Random Forest Classifier and Regressor.
   • Select independent variables and a dependent variable for training.
   • Enter values to predict outcomes using the trained model.
4. Feature Extraction:
   • Select columns to delete.
   • Analyze feature importance based on the selected target feature.
5. Explanation:
   • View a detailed explanation of why Random Forest models are robust to various data issues.

Contact

For any inquiries or issues, please contact:

• Phone: 9361381816
• Email: sankarkarthikeyan066@gmail.com
• LinkedIn: Sankar Karthikeyan
• Location: Coimbatore, India