Project / 01
Digital Behaviors
A mobile app that looks at device-use patterns and turns them into a clearer view of daily habits.
At a glance
- Android app
- Machine learning workflow
- Focused on everyday use patterns
Overview
This project started as an attempt to make digital habits easier to understand. Instead of looking at screen time as a single number, the app tries to capture a broader view of how a device is used during the day.
Problem
It is easy to lose track of how time is spent on a phone. Patterns can feel invisible until they are shown clearly, and that makes it harder to reflect on daily behavior.
My Approach
I focused on building a simple experience that could show meaningful patterns without becoming overly complex. The app combines simple data collection with machine learning so the output feels useful rather than abstract.
Technologies Used
- Android
- Python
- TensorFlow Lite
- Scikit-learn
- Random Forest
Architecture
Challenges
- Keeping the workflow understandable for a small project while still using ML meaningfully
- Designing the experience so the output felt useful and not over-engineered
- Balancing privacy considerations with data collection
What I Learned
- How to connect a mobile interface to a simple machine learning pipeline
- How to keep experimentation grounded in a practical use case
- How much clarity can come from a focused feature set
Future Improvements
- Refine the model with more realistic input data
- Improve the UI so insights are easier to read at a glance
- Explore a more complete local-first workflow
GitHub
The project source is available on GitHub for review and further exploration.