Inspiration
Shopping is grueling work. Be it games, hardware, office supplies - you'll either spending hours looking at dozens of tabs and evaluating how the products you find measure up, or you'll miss out on the deal that's best for you. We were inspired by challenges from Noibu and Cohere.
What it does
Shop Savant is an AI companion browser extension that extracts, analyzes and compare marketplace information for products you're interested in. It uses Natural Language Processing along with more conventional data-scraping techniques to extract the most relevant product attributes and to make guesses as to how they should be weighted and visualized. If the weighing doesn't align with your needs, you can always open comparison details and adjust it accordingly.
How we built it
We started simple by experimenting on the Co:here playground. We wanted to see whether NLP was able to compare products and tell us which one was better, and to our surprise, it did! From there we started planning from the bottom up: how the popup would look and behave, the connections between the different services we wanted to incorporate, and find a framework we could work with. We settled with using Angular for our extension and split the work into frontend and backend. Frontend was in charge of ensuring everything looked great and connecting moving pieces while backend created APIs and scraped data from the web. Once it was mostly done, backend was able to deliver the necessary data and frontend was able to present it in the extension.
Challenges we ran into
As with all coding projects, getting things to properly work (especially within such a short period of time). Our biggest nemesis during the project was JavaScript and its insistence on not allowing non-modules to import functions from other files. After hours of trying, we found a workaround that did the job and moved forward to create a mock we could demo.
Accomplishments that we're proud of
We came really far with this project, a lot further than we initially expected. First of all, we're happy with our idea and find that it can be very useful for people, especially when online shopping. We're proud of all of the new technologies that we've learned and how well we were able to work as a team.
What we learned
We learned A LOT. Nobody on the team knew about building web extensions beforehand or anything about AI and yet we somehow managed to do both. None of us were in our comfort zones with this new technology and made large strides within the fields of NLP, extensions, APIs, and Angular.
What's next for Shop-Savant
The next steps for Shop-Savant would be to expand to more sites than just Amazon and add cool new features like having NLP compare the attributes and add weights to them based off of what is important to the type of product - like how RAM and memory are important attributes in a computer but camera quality may not necessarily be.
Log in or sign up for Devpost to join the conversation.