RedPandas

About

For the last few years, we've noticed that data science is rapidly becoming a necessity in collegiate research even in non-technical fields - whether it's a bio lab or a sociology research group, students across the board are increasingly being expected to be to use and manipulate data without formal technical training. We noticed that even when the data-science tools themselves are understood, more technical operational requirements - such as cloud hosting and model deployment, are a major bottlenecks in the efficacy of research groups, as they often struggle to move projects past the local filesystem.

Red Pandas is a platform that makes data accessible to all. Users can import data, manipulate/visualize it, and share their results all from an easy-to-use web-based platform. We also integrated a GPT-based code generator in our text editor so that even those with no programming experience can use the platform as well.

Use Cases

• Hosts data science workflows without users ever having to interact with a cloud provider!
• Enables those with experience to write python and manipulate data from an easy-to-use code editor
• Novices can build code using GPT-3 without formal programming experience
• Can share entire projects (data + scripts), unlike any other platform.
• Iterative programming experiences that minimizes the monotony of Jupyter notebooks.

How we Built It

• CLI in Go for uploading even the largest datasets
• Frontend in Next.js React
• Interpreter in a Rust docker image that executes the Python provided by a user

Challenges We Ran Into

• Rust's (lack of) interoperability and the many many challenges that came with that
• Poor documentation for many of the libraries we used
• Maintaining good security-standards throughout the development process
• Assigning and delegating tasks efficiently so that our final product would be finished in time
• Docker build times
• General doziness

Accomplishments That We're Proud Of

• Teaching 2 team members Go entirely during the hackathon
• Architecting and decomposing a multi-faceted solution under a significant time constraint.
• Our journey as teammates and friends over the course of HoyaHacks

What We Learned

• Go
• How to Deploy Docker images to GCP

What's Next for Red Pandas

• Better User-Interface
• Real-Time Collaboration
• Report Creation
• Performance improvements