🏃‍♂️ Activity Recognition App

A modern Android application for collecting, visualizing, and analyzing human activity data.

🌟 Overview

This app is designed to help researchers and developers collect high-quality sensor data (Accelerometer & Gyroscope) for building Activity Recognition (HAR) models. It features real-time visualization, on-device classification, and robust data management.

✨ Key Features

1. 📊 Real-time Visualization

• Dynamic Line Charts: View live accelerometer and gyroscope data streams on scrolling line charts.
• Instant Feedback: See exactly what the sensors are "seeing" as you move.

2. 🧠 Real-time Classification

• Smart Detection: The app analyzes your movement in real-time to detect if you are ACTIVE (Walking, Running) or STATIONARY (Sitting, Standing).
• Heuristic Engine: Powered by a lightweight on-device algorithm.

3. 📂 Data Management

• Recording History: Browse all your past recording sessions in a clean list.
• Easy Sharing: Export your CSV data files to Google Drive, Email, or your computer with a single tap.
• Training vs Testing: Tag your data explicitly for Machine Learning workflows.

4. 📈 Session Statistics

• Insights: View aggregate statistics of your usage.
• Breakdown: See how much time you've spent recording each activity type.

🛠️ Tech Stack

• Language: Kotlin
• UI Framework: Jetpack Compose (Material 3)
• Navigation: Navigation Compose
• Architecture: MVVM-inspired pattern
• Sensors: Android SensorManager API

🚀 Getting Started

1. Clone the repo:

   git clone https://github.com/yourusername/activity-recognition.git

2. Open in Android Studio:
   • Select the project folder.
   • Wait for Gradle sync to complete.
3. Run:
   • Connect your Android device or start an Emulator.
   • Click Run (▶️).

📱 How to Use

1. Select Activity: Choose what you are doing (e.g., "WALKING") from the dropdown.
2. Start Recording: Tap the big "Start Recording" button.
3. Move: Perform the activity. You'll see "Detected: ACTIVE" if you are moving!
4. Stop & Save: Tap "Stop". The data is automatically saved.
5. Export: Go to the History tab to share the CSV file.

🤖 Machine Learning Workflow

Want to build your own model?

1. Collect data for different activities (Walking, Sitting, etc.).
2. Export the CSV files.
3. Use Python (Pandas/Scikit-learn) to train a classifier.
4. (Future) Integrate your TFLite model back into this app!

---

Built with ❤️ for the Open Source Community