We were inspired by a passion to travel. JetBlue airlines offered us a chance to figure out how to be critical flyers by understanding the massive amounts of data that is available about people's opinions on airlines.
What it does
Our app displays multiple different platforms where public sentiment can be found about JetBlue airlines. Specifically we analyzed travel review sites, Facebook and Twitter. Then the app allows the user to select the platform they want to learn about and it displays what the average rating of JetBlue was on a five point scale.
How we built it
Back-End Web Scraping
We used UiPath Studio to develop a bots that scrape each of the platforms mentioned. The travel review sites bot takes the star rating and number of users that reviewed from multiple travel sites. The Twitter bot takes tweets from #jetblue. And the Facebook bot takes posts from #jetblue. All of them store about 100 elements to a text file.
We first used a java to implement object oriented programming that processes the text files take from the bots. For the travel review sites it takes the weighted average of all the star ratings. This means that it accounts for the number of users that voted for each rating. The tweets and Facebook posts a value was assigned based on the overall sentiment of the post/tweet. We used data from a movie reviews file from Rotten Tomatoes to teach our program what type of words are associated with good or bad reviews. Then we assigned each platform an overall rating out of five.
We ended up converting the program to c# so that it can be used in visual studio.
Using Unity and Visual Studio we developed an Android app that acts as the user interface. It takes in the ratings for each platform and displays them for the user in an easy to view format.
Challenges we ran into
We had challenges with integrating every part of this project as there were multiple parts. It was also challenging converting our java code to c# and took a lot of debugging. We were all learning new languages and APIs so it was generally a large learning curve.
Accomplishments that we're proud of
We are happy that we were able to actually scrape data from off the web with an automated process. We were also to output the data in a form that was readable for the java program. And the data processing was able to provide an overall fairly reflective evaluation of user sentiment of JetBlue on each platform. In addition, getting the app up an working and displaying the data was a huge achievement.
What we learned
We learned how to use c# in unity, visual studio and UiPath studio.
What's next for CrowdMotions
In the future we will most likely incorporate an online database to store the data from the web scraping that can be downloaded from the app. We would also have the option to select other companies in addition to JetBlue.