JetBlueByMe

When faced with endless amount of feedback thrown at you through pages like Twitter and Trip Adviser.What do you do? We at JetBlueByMe have developed an algorithm that of what people are thinking.

Comment

Inspiration

We got our inspiration from looking at the tools provided to us in the Hackathon. We saw that we cold use the Google API’s effectively when analyzing the sentiment of the customers review on social media platforms. With the wide range of possibilities, it gave us we got the idea of using programs to see the data visually

What it does

JetBlueByMe is a program which takes over 16000 reviews from trip advisor, and hundreds of tweets from twitter to present them in a graphable way. The first representation is an effective yet simple word cloud which shows more frequently described adjective larger. The other is a bar graph to show which word appears most consistently.

How we built it

The first step was to scrape data off multiple websites. To do this a web scraping robot by UiPath was used. This saved a lot of time and allowed us to focus on other aspects of the program. For Twitter, Python had to be used in junction with Beautiful Soup library to extract the tweets and hashtags. This was only possible after receiving permission 10 hours after applying to Twitter for its API use. The Google sentiment API and Syntax API were used to create the final product. The syntax API helped extract the adjectives from the reviews so we can show a word cloud. To display the word cloud, the programming was done in R as it is an effective language for data manipulation.

Challenges we ran into

We were unable to initially use UiPath for Twitter to scrape data as it didn’t have a next button, so the robot did not continue on its own. This was fixed using beautiful soup on Python. Also, when trying to extract the adjectives, the compiling was very slow causing us to fall back about 2 hours. None of us knew the inns and outs of web hence it was a challenging problem for us.

Accomplishments that we're proud of

We are happy about finding an effective way to word scrape using both UiPath and BeautifulSoup. Also, we weren't aware that Google provided an API for sentiment analysis, access to that was a big plus. We learned how to utilize our tools and incorporated them into our project. We also used Firebase to help store data on the cloud so we know its secure.

What we learned

Word scraping was a big thing that we all learned as it was new to all of us. We had to extensively research before applying any idea. Most of the group did not know how to use the language R but we understood the basics by the end. We also learned how to set up a firebase and google-cloud service that will definitely be a big asset in our future programming endeavours.

What's next for JetBlueByMe

Our web scraping application can be optimized and we plan on getting a live feed set up to show reviews sentiment in real-time. With time and resources, we would be able to implement that.

Built With

Share this project:

Updates