Paper Scraper | Devpost

Logo
HomePage
500 Page Textbook being Parsed
Word Map of my resume

Inspiration

The whole concept of flipping through pages doesn't do many people justice, we wanted to go above and beyond of what a book can provide. We wanted to create a platform which could mine all the keywords associated with the thing you desired to learn more about. We wanted something that you could play around with and something that works for all books out there.

What it does

It does, what we expected it to do. Give us insight and a quick overview of any book that we might utilize in our courses or simply for pleasure. It takes any PDF file and gives you a WordCloud of all the Keywords that are frequently used in the book. We used simple constructs of Natural Language Processing to extract these insights for the user to gain more knowledge in a quick span. We wanted something interactive and fun for our users. It also allows you to search for a key phrase that you might be looking for in a book and gives you a visual representation of how many times it is been referred to. An excellent way to see what books are relevant for what you desire to learn.

How we built it

For the front-end, we utilized AngularJS, HTML, CSS, and Javascript. For our back-end, we utilized the NodeJS and ExpressJS framework. We are hosting the web application using the Google Cloud Platform.

Challenges we ran into

UI design challenges, focusing on what is the best layout for overall great user experience. Followed with many grueling hours of Mining PDF files and identification of the keywords.

Accomplishments that we're proud of

The overall look and feel of the product, next to no bugs, portability and overall a great learning experience.

What we learned

We had a couple of first-timers, a learned a great deal about web development, utilizing version control systems. And I think we all had a fun time debating how to visualize the large amounts of data being parsed.

What's next for Smash Books

Migrating this platform to mobile and desktop applications would have been pretty amazing to have. Improving upon the approach we took on mining the data and incorporating multi-layered associations in a graphical form. Mining the web to provide more insight into the content.