BettMAT, a media analysis chatbot
Upload a spreadsheet of tech media coverage and ask questions about it. The chatbot answers from your data and nothing else.
- Year
- 2024
- Tools
- Python, OpenAI API, Gradio, pandas
The aim
Build a chatbot that analyses and summarises tech media content uploaded by the user. It must only work on the data provided and only answer questions about that data.
How it works
Users upload a CSV, Google Sheet or Excel export that follows a provided template. The data feeds a GPT model through the OpenAI API, and a simple Gradio chat interface lets users ask things like “what are the common themes this quarter?”. If a question is out of scope, the model says so instead of guessing.
What we delivered
- A working chatbot with summarisation and question-and-answer features.
- A pipeline for uploading, parsing and analysing media content.
- Data templates, plus a user guide for deployment and use.
The outcome
Teams can summarise their media coverage in minutes and spot the trends in it, without anyone reading every article.