High-Level Project Summary
Our project is a web browser that help users to do resources searching. The main goal is to save users from burdens of filtering large amounts of articles & getting relative information. In our website, upload records articles from the public NTRS, you can grep keywords, outline and the picture showing words relationships by technique AI. Then click in slice of keyword pie-chart just show, you can access more highly-relative information. Moreover, you can search your own keyword with the search bar in the header. This Browser can offer users more effective ways to view articles quickly and search articles wisely !
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Our project name is PDF QUICK VIEWER, users can easily take a quick view to papers. We make visualized charts which show keywords, and hyperlinks to relative documents.
Here are the benefits of our project.
- Reduce a lot of time in reading NTRS(or others) paper
- Easy to find the correlation between papers by clicking corresponding keywords
- Create convenience to researchers
Hope this project will bring convenience to people in their studies
In the future, hope this quick search applications can help people to manage and search files in their laptops or even database for big business
We use:
Fastapi for backend, Canva for logo
Space Agency Data
We used NTRS documents for our demonstration, and chose the label "space and planetary science" for example.
There were about tens of thousands of documents with this label, so we decided to make classification and quick review to them.
The result showed that our keywords list was reliable to NTRS documents. It inspired our project, because we could have more reliability on our presentation.
Hackathon Journey
This was our first participation on NASA Space Apps Challenge.
The reason why we chose this challenge is all of the members are doing study on nlp, we tried to perform an user-friendly platform to read documents.
To achieve this goal, we have three main tasks:
- text-summarization by using bertsumext
- keywords list by using keybert
- front/back end setup by using fastapi.
What really help us was the work division, it enable us to identify clearly to our task, made our work of integration easier.
We were extremely honored to participate in this challenge, it was indeed fun.
References
Tags
#NLP, #search, #filter, #NLPDOG