Inspiration
Smaller businesses have to be very careful with how they spend their resources because each dollar spent is incredibly valuable. Kuttt hopes to makes that decision easier, by compiling reviews together so small businesses know where they need to improve. Often times, reading more reviews makes things more and more confusing and Kuttt summarizes the key information into a digestible format.
What it does
Kuttt evaluates collections of reviews and performs sentiment analysis to categorize positive and negative reviews. From there, the results are mapped into positive and negative word clouds to share the most common praises and complaints users have with the product at hand. Our computed results are displayed on our website for users to get an accurate summary of reviews.
How we built it
On the back-end, we used a dataset to develop training data for our model. A review's rating allowed us to superviseFrom there our model categorized good and bad reviews and compiled each into their own pools. Our results were transmitted to our website where word clouds display the most common positive and negative thoughts about the product.
Challenges we ran into
We used many different interfaces to program collaboratively and compiling all our work into one final product was difficult. We used Python, HTML, CSS, JavaScript, and more, and combining our computation and website setup was difficult. We also had a lot of difficulty in building our website and making it look the way we wanted it to. Options like displaying popups were a huge problem and something we had to bail on in the end.
Accomplishments that we're proud of
We're really proud of the fact that we were able to properly link all of our code together into a final product. We're all beginners and it was our first time working together, and we're super happy with the fact that we connected our collaborative Python notebook with a working HTML/CSS based website.
What we learned
We learned a lot about how sentiment analysis models work. Additionally, we learned about how to collaboratively code together and link our work into a final product.
What's next for Kuttt
We want to make Kuttt a little more meaningful for a vendor and small businesses. By looking over reviews of their own product, they can see the biggest complaints about their product but also the biggest selling points. With this information, a company will have a better idea about where to direct their resources and how to improve. Finally, we want to create a filtration and search system where clicking on a word in the word cloud will pull up all reviews mentioning that word if a user would like to take a deeper dive.
Log in or sign up for Devpost to join the conversation.