Fitness App Step Counter

This project is part of the 'Enterprise Guide to DataOps and MLOps' lab assignment. It involves building a Python-based step counter that tracks daily activity and provides insights on weekly performance, along with predictive analytics using AutoML techniques to forecast future step counts.

Features

StepCounter Class

The core of this project is the StepCounter class, which tracks the following metrics:

• Daily Steps: Logs the number of steps taken each day.
• Total Weekly Steps: Calculates the sum of steps over the past week.
• Average Daily Steps: Provides the average number of daily steps over a week.
• Activity Level Feedback: Categorizes activity as low, moderate, or high based on the total number of steps.

The data is stored in a table with the following columns:

• Date: The date on which the steps were taken.
• Steps: The number of steps taken.
• Calories_Burned: Estimated calories burned based on the number of steps.

Project Structure

This project follows a structured branching strategy:

• Main branch: Holds stable and tested versions of the project.
• Development branch: Used for integrating new features like setting daily step goals.
• Feature branches: Separate branches for specific functionalities, such as:
  • Calories Calculation: Implement functionality to calculate calories burned from steps.
  • Other features: Additional functionality is merged into development, and later into main.

AutoML Predictions

The project uses AutoML to predict future step counts based on historical activity data. Three models are compared:

1. Linear Regression
2. KNeighbors Regressor
3. Lasso

The model with the lowest RMSE (Root Mean Squared Error) is selected for future predictions.

Columns used for prediction:

• Age
• Steps
• Calories_Burned
• Weight

Testing

We use unittest to ensure the accuracy and reliability of the StepCounter class:

• Verify total weekly steps calculation.
• Ensure high activity levels are correctly identified.

Installation

To set up the project locally:

1. Clone the repository:

   git clone https://github.com/Pratikchetry/fitness-app-step-counter.git

2. Navigate to the project directory:

   cd fitness-app-step-counter

3. Install the necessary packages:

   pip install -r requirements.txt

Usage

1. Initialize the StepCounter class to track daily steps:

   from step_counter import StepCounter
   
   counter = StepCounter()

2. Log daily steps and get weekly summaries:

   counter.add_daily_steps("2024-09-01", 10000)
   total_steps = counter.get_weekly_steps()

3. Predict future step counts:

   from auto_ml import predict_future_steps
   prediction = predict_future_steps(user_data)

Branching Strategy

• Main Branch: Stable versions of the project.
• Development Branch: Ongoing work and feature integration.
• Feature Branches: New features such as daily goals or calories burned, merged back into development before being integrated into main.