The Breakdown of the Tweets

Users wrote tweets in a total of twelve languages: ar (Arabic), ca (Canadian French), de (German), en (English), en-gb (British English), es (Spanish), fa (Farsi), fi (Finnish), fr (French), it (Italian), ja (Japanese), ko Korean), nl (Dutch), pt (Portuguese), ru (Russian), sv (Swedish), th (Thai), tr (Turkish).

19,328 tweets were written in English. I used a total of 21,072 tweets. Therefore, 19,328/21,072 = 91.72% of the tweets were written in English.

I found that 1.57 percent of tweets were written in Spanish (330/21,072) and 0.31 percent of tweets were written in Arabic (65/21,072).

President Trump’s travel ban on seven Muslim countries took effect on Saturday, January 28th, but due to the weekend, the media did not cover it intensely until late Saturday into early Sunday. #DeleteUber, which stemmed from Uber eliminating surge pricing for those who travelled to JFK Airport, trended on Sunday, January 29th and Monday, January 30th. I believe that people did not tweet as much on #nobannowall on Wednesday, February 1st because people were still trying to understand the implications of the travel ban. Towards the end of the week, people started posting individual stories of nationals from the countries with valid U.S. visas stuck in their home countries or in U.S. airports and not allowed to return to the U.S. Outrage followed with numerous protests by around Wednesday, February 1st to the beginning of the next week, ending at around Tuesday, February 7th.

According to Google News, on February 1st, people started to express outrage on the executive order through smaller protests concentrated around liberal centers of thoughts such as universities. Students from large universities such as Arizona State and Utah State rallied at their prospective colleges, but the protests did not expand outside of places with lots of known activism. Celebrities commented negatively on the travel ban on their social media feeds. On February 2nd, a spike in tweets happened probably because more national, large-scale protests occurred. More than 1,200 Comcast employees walked out of the job in protest of the Executive Order. Thousands of people marched in New York. Google News did not cover any new protests or well-known commentators talking about the Executive Order on February 3rd. Instead, media outlets summarized the events (including protests, commentary, Trump’s own comments) throughout the week since it was a Friday. Tweets spiked again on Saturday, February 4th after the media focused on protests abroad. They referenced similarities with the Women’s March on January 21st and expressed concerned about the U.S.’s international reputation.

Below are the stats of my data set:

Compared to the entire class, I ranked second in both the mean and median number of tweets, behind only Taylor’s #climatechange. My data set did not have a mode and only two other people in the class had a mode. I browsed through other classmates’ graphs and noticed that they selected a wider range of dates for the tweets, meaning that they more likely had days where their hashtag did not trend. In contrast, not only did I pick a shorter timeframe, but a timeframe where my hashtag peaked and trended more than any other time during the semester. I choose this timeframe because I thought it would be most interested to learn about how people used the hashtag as close to when President Trump announced the Executive Order. Referring to Gillespie’s discussion of the trending algorithm, I believe that my tweet probably did not trend until February 2nd when the number of tweets increased more than half. However, I did feel that people used my hashtag starting the weekend. I doubt that Taylor’s hashtag ever trends since individuals probably continuously use the hashtag and there are no “spikes” in usage.

I ranked third in both Min and Max, and Range. I ranked fifth in count. This means that I had more daily tweets that other classmates, but a shorter timeframe than them. #NoDAPL, #ClimateChange, and #MakeAmericaGreatAgain, and #BlackLivesMatter also had significant number of tweets per day and either outperformed or within range of my hashtag. I am not particularly sure if these tweets “trended” during the same timeframe, but I noticed that all of them ranked similarly at the top of the Min, Max, and Range columns. I am interested in learning if all of the hashtags trended or some did not receive much attention on the trending list after reading Gillespie’s discussion on how some “popular” tweets do not trend.

I expected to see more tweets in other languages, particularly in Arabic since the Executive Order targeted Arab speaking nations and the majority of Muslims speak Arabic. However, since less than 0.3 percent of my dataset wrote in Arabic, it either shows that Arabic-speaking Muslims are marginalized on Twitter or many people outside of the English speaking world do not have access to Twitter. The most common languages such as Chinese are not even represented in my dataset. I am not sure if my dataset truly represents people are affected by the Executive Order or people who are not affected at all, but only show solidarity through a social media post.

Works Cited

Gillespie, Tarleton. 2012. “Can an Algorithm Be Wrong?” Limn (2).