While we were researching for history class, we realized that a tool that could markdown important information on sites would be useful. This would be better than taking down notes on a separate page or tab. This idea expanded into a full-fledged academic research tool.
What it does
Spade is a chrome extension that simplifies research. Spade is a lightweight, full-featured research tool. It allows users to easily markup pages and save annotations, review source credibility, find answers, and copy generated citations. It solves the problem of procrastination in research and time efficiency because note-taking is extremely tedious and with this tool, students can easily find good sources and mark them up, getting quick answers to their questions.
How we built it
We utilized p5.js to enable users to annotate pages. We saved these annotations on locally on chrome so users can always review them. Using a Tensorflow model, we created a reliability rating system that gives rankings on the url, bias, relevance, grammar, and credibility. This scoring algorithm also normalizes numbers such as grammar mistakes and dates with an exponential function to make the report readable. We used a Transformers question answering service pipeline so users can easily find answers in their sources. To create citations, we used Beautiful Soup to webscrape the required information.
Challenges we ran into
One of the biggest challenges we ran into was figuring out how to not max out the memory limits on Heroku. It was difficult figuring out how to split up corpora and efficiently process them for our question answering service. We also had a problem with utilizing p5js as there wasn’t much documentation on how to use its features easily with a chrome extension, and we had also never added a backend to our chrome extensions before. The biggest mathematical challenge was weighing the credibility of a site based on criteria as we had to normalize the raw scores that could range from 0-infinity and fit them within a 1-10 range. We had to graph a curve that fit our needs specifically that wasn't linear so that we could demonstrate the particular point that the credibility of the site would drop a lot.
What's next for Spade
We would like to add in more citation formats. Currently, Spade only cites in APA. We would also like to make our reliability ratings more accurate. We would also like to link a cloud-based storage system that users can save annotations and webpage citations on. This would make the same annotations accessible across multiple devices for users. Our question answering service also takes a long time to complete. Making this process faster would be prime. We would to put this on the web store and reach students all over the country to take notes by annotating sites instead of going back and forth between screens!