MapIT

Turn your notes into a mind map

Comment

🏝️ Inspiration

Something super uncontroversial is that note-taking is pretty boring. Sometimes, it’s not even very effective because you don’t even think about what you’re writing down. We just do it because we have to. To take effective notes, though, the note-taker must be engaged and thinking critically, and a core part of critical thinking is making connections. Mind-maps are super useful for making these connections, but they take a lot of effort to make. That’s where MapIT comes in! With MapIT, students (or anybody else!) can transform their notes into mind-maps with a single click! MapIT just helps students make their notes more meaningful, teaching them how to learn more effectively.

🔖 How it works

In short, what our application basically does is transform notes into mind-maps. How this is achieved is explained more in detail below:

  1. The user first uploads an image file
  2. The image interpreted by an AI which extracts the text
  3. The text is then categorized by another AI, and formatted into Markdown.
  4. We then use MarkMap to display the resulting Markdown as a mindmap

✏️ How we built it

  • Frontend: the UI was built using React & CSS modules
  • Backend: Python, Flask, and Tensorflow were used to run the ML Models and we hosted our API endpoints on Heroku
  • Machine learning: We used Azure Computer Vision for handwriting recognition & the Universal Sentence Encoder for text categorization
  • Design: Our designs were created using Adobe Illustrator, Adobe Photoshop, & Figma

🚨 Challenges we ran into

  • Organizing and processing data is A LOT harder than we thought
  • Figuring out how MarkMap worked was also a challenge because it wasn’t a very common library
  • We were also working in different time zones so it was difficult making sure everyone was on the same wave-length
  • repl.it has storage limitations and we couldn't download a model there so we had to move to Google Colab.

🤩 Accomplishments that we're proud of

  • We’re just super proud of how everything turned out, and how we managed to connect everything together!

📚 What we learned

  • Tensorflow models and Natural Language Processing!
  • How to implement next.js as the frontend and flask as the backend.
  • Data cleaning and preprocessing/postprocessing take up a significant amount of time!

💭 What's next for MapIT

  • We plan on finding out how to make the program connect concepts together and extract those keywords more accurately.
  • We also plan to add a way for their user to paste their text in directly

Built With

Share this project:

Updates