Airlines U.S. Mode of Transportation on Twitter
Project GitHub | LinkedIn: Niki Moritz Hao-Wei Matthew Oren
The skills we demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Introduction
As technology grows and globalization develops, more and more consumers are using airplanes as their mode of transportation. With so many choices when it comes to flying, and so many airlines in the U.S. competing with each other, it is valuable to analyze the feedback given by consumers on Twitter. This study therefore takes this data set from Kaggle, with more than 14,000 tweets during the month of February 2015 over 6 major U.S. airlines. This data includes an overall rating, which can be negative, positive or neutral, along with their comments. The RShiny visualization can be viewed at https://fredzeng.shinyapps.io/shinyDashBoard/ and the code at https://github.com/freddy90503/ShinyProject.
Overall
First I created a graph that shows the difference in quantity of reviews by three levels of rating. We can see that most of the tweets were negative. After that, I created two graphs to visualize how each individual airline performs. We can see that United Airlines has the most reviews but also has the highest negative percentage rate, while Virgin America has the fewest reviews but also the lowest negative rate.



By Airlines
Next, I conducted an analysis on each airline over three different charts in order to visualize passengers' reviews from three different perspectives.



By Words
After that, I aggregated keywords from each airline's comments to visualize the key reasons people didn't enjoy their experience. First I conducted a word count for each airline, and then built a heat map, followed by word clouds of the negative reviews.







By Location
Following that, I visualized the aggregation of IP addresses that came with the data set with a map to identify where users are located. We can see that the majority of users are from North America, some are from Europe, and a small amount from Southeast Asia.

Conclusion
With this project, I was able to gain some key insights:
- Most Internet users did not enjoy their experiences with major US airlines.
- The best airline of the six is Virgin America, and the worst is United Airlines.
- The top three reasons for negative reviews are bad customer service, delays, and cancellations.
The last insight is particularly important, as to improve ratings and customer retention, I would advise airlines to focus on identifying why poor service, delays, and cancellations occur. Indeed, with more time, that is where I would take this project: I would expand my data set and perform statistical regressions to help narrow down those root causes.