Insightia

• 

  home page

Inspiration

I was inspired by how overwhelming raw datasets can feel, especially for people without a technical background. Many users have data but lack the tools or skills to quickly understand it. I wanted to create something that feels like having a data analyst on demand — a system that can instantly interpret a dataset and explain it in plain language.

What it does

Insightia is an AI-powered data analyst copilot. Users upload a dataset and immediately receive automated profiling, including structure, missing values, and statistical patterns. It then generates a professional AI report and allows users to ask questions about their data in natural language, turning raw numbers into clear, actionable insights.

How we built it

I built Insightia using Python and FastAPI for the backend, with Pandas handling automated data profiling and statistical analysis. These insights are passed to the Gemini API, which generates structured reports and answers user questions. The frontend provides a simple interface for uploading datasets, viewing analysis, and interacting with the AI chatbot.

Challenges we ran into

One major challenge was controlling how the AI responded. Large language models tend to be overly verbose or make assumptions about the dataset’s theme. I had to carefully engineer prompts to ensure responses stayed professional, concise, and strictly data-driven. Handling messy real-world data, such as inconsistent column names and encoding issues, was another key hurdle.

Accomplishments that we're proud of

One major challenge was controlling how the AI responded. Large language models tend to be overly verbose or make assumptions about the dataset’s theme. I had to carefully engineer prompts to ensure responses stayed professional, concise, and strictly data-driven. Handling messy real-world data, such as inconsistent column names and encoding issues, was another key hurdle.

What we learned

I learned how to bridge traditional data analysis with AI-generated explanations. Prompt design turned out to be just as important as the code itself, especially when ensuring reliable, domain-neutral outputs. I also gained experience in designing systems that are both technically robust and user-friendly.

What's next for Insightia

Next, I plan to expand file format support, add visualizations for detected patterns, and introduce deeper statistical analysis such as correlations and trend detection. I also want to improve context memory so the AI can reference previous questions, making Insightia feel even more like a true data analysis partner.

Built With

• css3
• fastapi
• google-gemini-api(3-flash)
• html5
• javascript
• pandas
• python
• uvicorn