Inspiration
We believe that a company should employ a customer-first approach when developing products and solutions. A great customer experience promotes customer loyalty and better products/services overall. Learning that applying natural language processing to whole datasets of customer feedback could result in categorization into feedback as positive or negative using sentiment analysis, we were inspired to find a way to make improving customer experience for businesses and teams as smooth as possible!
What it does
From various data inputs (whether it be from customer reviews, comments on social media, or recorded calls from contact support), we apply a sentiment analysis on dataset of reviews and assign each with a score on how positive or negative it is, allowing us to classify reviews as positive or negative. Then, we find the top few positive and negative reviews and aggregate the top keywords in order to figure out the specific product or specific area (response time, affordability, branding, quality, ease of use, customer service) a company needs to improve. Then, we send this data to a frontend page in order to display a data visualization in order to figure out the best course of action to improve a product/service.
How we built it
Frontend: React Backend: NodeJS, Express, Python Sentiment Analysis is done using the python library NLTK Twitter web scrapping is done using python library twint
Challenges we ran into
Though we could scrap tweets from Twitter and perform sentiment analysis on this dataset of tweets, we had challenges connecting the frontend and backend together.
Accomplishments that we're proud of
We are proud that we have a frontend client webpage, a backend client, and a usable python script for web scrapping and sentiment analysis. We are also proud of being able to send simple CRUD HTTP requests to and from the front and back ends.
What we learned
As this is our first time doing full-stack development, we learned a great deal about the tools and methods that go into making a full-stack product. From using React to NodeJS and Express to various Python libraries, we learned the process of full-stack development.
What's next for CBRE-Challenge
Classification of data into categories and specific products/services in order to be able to hone in on specific areas of improvement. Connecting the frontend and backend. Data visualization.
Log in or sign up for Devpost to join the conversation.