Inspiration
We noticed that while many data tools exist, they often fall short. They either require deep technical knowledge, lack flexibility, or can't handle the complexity and nuances of real-world, messy data. Our inspiration was to build a tool that bridges this gap. A solution powerful enough for complex tasks and intuitive enough for anyone to use.
What it does
- Chunks up your data into manageable sizes where the models will perform at maximal accuracy
- Analyzes the individual chunks in order to identify all the issues and problems that will require cleaning in your files and then combines them
- Based on the final combined evaluation, runs data cleaning tools on the data or if one doesnt exist, creates a tool to handle those use cases
- Presents the final cleaned data to you in a downloadable CSV file ready to go
How we built it
- Analysed other available tools on the internet and realised where they fell short
- Tackled the most common problems
Challenges we ran into
- Data being improperly chunked
- Incorrect specification for tool generation ## Accomplishments that we're proud of
- Got a finished product ## What we learned
- How to get agents to talk to each other
- How to direct agents and get them to produce what you want ## What's next for Scrub Hub
- Tackling reliability issues as well as decrease file cleaning times
Log in or sign up for Devpost to join the conversation.