Data Generator

A Python tool that generates new, realistic data from a small existing dataset, making it possible to train AI models when real data is limited.

About the project

Training an AI model usually requires a lot of data, but in many real situations only a small sample is available. This project is a data generator that takes an existing dataset and extends it with new, synthetic data points that follow the same patterns as the original. The tool has a user interface where the user can load a dataset, choose how much data to generate, and export the result.

Challenges and solutions

A key challenge was making sure the generated data was actually useful: realistic enough to reflect the original, but varied enough to add new information instead of just copying it. I experimented with both AI-based generation and more traditional algorithmic methods.

What I learned

Besides developing my Python skills, the biggest lesson was about when to use AI and when not to. It's tempting to add AI to everything, but a well-designed algorithm is often faster, cheaper and easier to understand and test. I now think of AI as one tool among many, to be used where it actually solves a problem better than simpler methods.