Inspiration

What it does

AnalyzeThis is a tool that helps business owners detect what the people around the internet are saying about their products. With this, they can easily figure out what they’re doing right or wrong and how they can improve with some metrics crafted to quickly convey customer sentiment and overall improve customer experience.

How we built it

We used NumPy and Python Pandas to gather data and display the metrics from a pre-existing dataset as an example. It classifies positive and negative reviews by looking at word count and word choice. We then marketed the service on a website built with Wix by showing different analyses the program can display to a developer to easily inform them of what changes they need to make on their application. There is a section on the website explaining how the code was made and how it works too.

Challenges we ran into

Execution of the code was taking a longer time than expected but that was likely due to the large dataset and a slow computer.

Accomplishments that we're proud of

The code searching for keywords and displaying a graph of the tone was a great accomplishment. Also the website interface looks modern and flashy.

What we learned

We learned how common and important the application of data science is and how to clearly identify a problem and develop a solution. Since most of us did not have much experience to start with, this Hackathon was great for fostering our imagination and led us to wanting to learn more backend to implement such solutions in the future.

What's next for AnalyzeThis

In the future, AnalyzeThis can be programmed to have more precision while classifying positive and negative reviews. There is still a margin of error due to imperfections and outliers. Also an increased vocabulary to search for keywords can be made with enough time.

A mobile app and more modes of viewing metrics such as different types of graphs or tables can also be made with time. Instead of having too many new features, the future would expand upon making what is already there more precise and accessible.

Built With

Share this project:

Updates