Inspiration
I wanted to build a project that blends my coding skills with data analytics, showing how machine learning models can be made interactive and accessible through a simple web interface. The goal was to create something demo‑ready for hackathons and portfolio showcases.
What it does
DataForge AI lets users load datasets, run predictions using a trained model, and view results instantly through a clean web dashboard. It bridges the gap between backend analytics and frontend usability.
How I built it
- Backend: Flask + Flask‑CORS for API routes, Pandas for data handling, Scikit‑learn for ML models, Joblib for persistence.
- Frontend: HTML, CSS, and JavaScript for the dashboard, with fetch calls to backend routes.
- Environment: Virtualenv for dependency isolation, VS Code for development, GitHub for version control.
Challenges I ran into
- Activating and managing Python virtual environments in PowerShell.
- Ensuring frontend and backend routes matched correctly.
- Handling JSON responses cleanly with proper Content‑Type headers.
Accomplishments that I'm proud of
- Built a working full‑stack app that connects ML predictions to a web UI.
- Debugged environment issues and got Flask running smoothly.
- Structured the project with professional polish (requirements.txt, venv, clear folder separation).
What I learned
- How to connect Flask APIs with a JavaScript frontend.
- Best practices for error handling and JSON responses.
- The importance of reproducible environments and clean documentation.
What's next for Untitled
- Add a
/trainendpoint to retrain models dynamically. - Improve error handling and input validation.
- Deploy the app online (Heroku, Render, or AWS) for public demos.
- Enhance the frontend with charts and visualizations.
Built With
- flask-cors
- html
- javascript
- joblib
- learn
- numpy
- pandas
- scikit