Pattern

Pattern is a native Android personal analytics application that helps users understand their daily routines through data.

Instead of simply recording habits, Pattern analyzes historical data to establish a personal baseline, identify unusual patterns, visualize trends, and discover relationships between different daily metrics.

Features

• Personal Analytics Dashboard — View your recent activity and overall Pattern Score.
• Daily Data Logging — Track sleep, screen time, study, exercise, mood, and productivity.
• Anomaly Detection — Identify values that are unusually different from your personal baseline.
• Trend Analysis — Visualize changes across 7, 14, and 30-day periods.
• Correlation Analysis — Explore relationships between metrics such as sleep and productivity.
• History — Review, edit, and delete previous daily records.
• Local Persistence — Data is stored locally using Room Database.
• Professional Dark UI — Built with a clean, modern Android interface.
• Statistical Testing — Core analytical functions include tests for edge cases and statistical accuracy.

How It Works

Pattern uses the user's historical data to create a personal baseline rather than relying on fixed population-wide thresholds.

For anomaly detection, the application uses statistical analysis based on the z-score:

z = |value - baseline mean| / standard deviation

The application classifies observations as:

Classification	Z-score
Normal	< 2.0
Unusual	2.0 – < 2.8
Very Unusual	≥ 2.8

When evaluating a particular observation, that observation is excluded from its own baseline calculation to avoid influencing its anomaly score.

When there is insufficient historical data, Pattern uses a percentage-deviation fallback.

📊 Metrics Tracked

Pattern currently analyzes:

• Sleep duration
• Screen time
• Study time
• Exercise duration
• Mood
• Productivity

These metrics can be viewed individually through charts and statistical summaries.

Technology Stack

• Kotlin
• Jetpack Compose
• Material 3
• Room Database
• Android Architecture Components
• ViewModel
• Kotlin Coroutines
• Jetpack Navigation
• Custom statistical analysis
• Data visualization

Architecture

Pattern follows a layered Android architecture:

┌──────────────────────────┐
│     Jetpack Compose UI   │
└────────────┬─────────────┘
             │
┌────────────▼─────────────┐
│       ViewModels         │
└────────────┬─────────────┘
             │
┌────────────▼─────────────┐
│       Repository         │
└────────────┬─────────────┘
             │
┌────────────▼─────────────┐
│       Room Database      │
└──────────────────────────┘

        Analytics Layer
              │
     ┌────────▼────────┐
     │ Statistical     │
     │ Analysis        │
     ├─────────────────┤
     │ Mean            │
     │ Median          │
     │ Std. Deviation  │
     │ Z-Score         │
     │ Correlation     │
     │ Anomaly Score   │
     └─────────────────┘

Getting Started

Requirements

• Android Studio
• JDK 17+
• Android SDK
• Android device or emulator

Installation

Clone the repository:

git clone https://github.com/YOUR_USERNAME/YOUR_REPOSITORY.git

Open the project in Android Studio and allow Gradle to synchronize.

Then run the application on an Android emulator or connected Android device.

🧪 Testing

The project includes tests for core statistical functionality and important edge cases, including:

• Mean and median calculations
• Standard deviation
• Z-score calculations
• Anomaly thresholds
• Baseline exclusion
• Zero standard deviation
• Insufficient historical data
• Pearson correlation
• Pattern Score boundaries

🔒 Privacy

Pattern is designed around local-first data storage.

Daily personal data is stored locally using Room Database and does not require a user account or cloud backend.

🎯 Project Purpose

Pattern was built as a practical exploration of:

• Native Android development
• Kotlin and Jetpack Compose
• Local database architecture
• Statistical data analysis
• Anomaly detection
• Data visualization
• Correlation analysis
• Clean and maintainable application architecture

The goal was to build a small application where data analysis is part of the core product, rather than simply creating another CRUD-based mobile application.

📌 Future Improvements

Potential future enhancements include:

• More advanced anomaly detection algorithms
• Exporting personal data as CSV
• Additional metrics
• Custom metric creation
• Longer-term trend analysis
• Optional cloud backup
• More advanced statistical models

📄 License

This project is available for educational and personal use.