Flappy Search: AI for Finding Datasets

Meet Flappy, your parrot pal! Flappy doesn’t mimic like other parrots--when Flappy hears an idea, Flappy echoes back with the perfect datasets just for you.

Comment

Inspiration

Flappy Search was inspired by several sources, including Kaggle, HuggingFace, and Reddit threads where users often ask for master lists of datasets. The idea also came from struggling with finding relevant datasets.

What it does

Flappy Search is a web app designed to assist students, researchers, and data enthusiasts in finding relevant datasets quickly and effortlessly. Searching through multiple dataset platforms with various keywords is often tedious and overwhelming. Flappy, your parrot pal, simplifies this process. By analyzing your project description, Flappy extracts keywords by using spacy and Tf-Idf Vectorizer, queries APIs with keywords to retrieve datasets, and uses cosine similarity analysis to deliver datasets most relevant to your project.

How we built it

  • Frontend: Built with React, utilizing CodeBuff for UI components, infinite scrolling, and page routing (e.g., navigation bar and dynamic dataset display).
  • Backend: Used Flask for keyword extraction and dataset querying, using Python's libraries for natural language processing and cosine similarity calculations.
  • Python Libraries/Tools: SpaCy, Tf-Idf Vectorizer, scikit-learn
  • Authentication: Implemented user authentication and hosting using Firebase.
  • Design: Designed prototypes in Figma

Challenges we ran into

  • Setting up Firebase hosting for the first time and navigating Firestore for data storage and updates.
  • Connecting Flask and JavaScript seamlessly to leverage Python APIs in a React-based frontend.
  • Managing updates to Firestore documents, particularly when handling lists of datasets.
  • Debugging issues with API integrations

Accomplishments that we're proud of

  • Successfully building an app that integrates frontend, backend, and design within a short timeline (my first hackathon!).
  • Creating an intuitive design inspired by a parrot theme, making the app functional and fun.
  • Mastering new tools and workflows, such as Firebase authentication and Flask-React integration, to create a seamless user experience.

What we learned

  • Gained hands-on experience with Firebase, particularly hosting and authentication workflows.
  • Learned how to bridge the gap between Flask and React, combining Python backend functionality with modern JavaScript frameworks.
  • Deepened understanding of dataset querying and cosine similarity analysis

What's next for Flappy Search: AI for Datasets

  • Advanced Search Filters: Implement filters for file formats, dataset size, and sorting options (popularity, relevance, date).
  • Expanded Dataset Sources: Add more APIs and live web scraping to increase the number of datasets available through Flappy Search.
  • Enhanced Models: Integrate advanced models for better keyword extraction and relevance prediction.
  • Public Deployment: Deploy the app to a wider audience to support students, researchers, and professionals globally.

Built With

Share this project:

Updates

Submission history