St. Lawrence University
Emily Balter

Question: How does filtering in users from the U.S. aged 18-24 and the use of videos/images on posts impact the keywords and content of tweets that have been posted from when Michael brown was shot (August 9, 2014) until a week later (August 16)?

The word cloud I created did help me see common themes in the tweets I was analyzing such as #ferguson, police, #mikebrown, protests, and #dontshoot. I was interested in seeing how younger activists and journalists reacted on twitter to the Michael Brown shooting, and what types of media they were posting during and throughout the week of the event. The most frequently used words are directly related to the images and videos used on tweets. Images include protestors on the streets of Ferguson, Michael Brown’s mother, tear gas, the site where Brown was killed, and looted stores. Many videos are real time scenes of the protests. Antonio French, a city alderman in St. Louis is one of the most active live time tweeters in the results. From looking at my word cloud, I was interesting in finding out when key words and hashtags were used most frequently throughout the week. I was interested in using key words to observe the progression of main events during the week of the shooting. The words “looting” and “store” were more frequent towards the end of the week due to protestors looting stores, specifically the one Michael Brown was thought to have stolen cigarettes from. Violent protests and police brutality were evident in the photos and videos posted, and therefore I was curious why violence and brutality were not key words. 

When constructing my word cloud, I decided to focus on photo and video posts during the week of the shooting because visuals often tell powerful stories of a timeline of events. Using photo and video for my word cloud helped me to understand the most important events immediately following the death of Michael Brown. Since most of the video and photos were live, I could see what was happening at an exact moment and who was posting these stories. Through using original posts, I was able to observe who was in Ferguson experiencing the real action. I chose the age range of 18-24 because I hypothesized that younger people would be the most likely to use media in their tweets. I was also interested in seeing the reactions of younger activists to the shooting. 

From my initial impressions of the search results, I assumed that two influential male news reporters dominated the tweets. I also noticed that many of the journalists and activists posting were likely older than age 24. If age was not included on a users profile, I assume that they were automatically included in the results. After looking more in depth at the results, although the majority of tweets were from reporters at the Huffington Post, CNN, or the New York Times, a few activists were involved and influential. I also noticed the rise of protests and forceful police presence throughout the week. This rising tension often resulted in violence. 

In my word cloud, it is clear that information about #ferguson, police, #mikebrown, and ferguson are emphasized since these words occur most frequently in my results. The word “right” also caught my attention. In many tweets, “right”, refers to “right now”, meaning happening in the moment. Therefore, there is an emphasis on real time events that are evident through photos and videos. The majority of the words surround the most frequent word #ferguson because these words are mentioned in a tweet with #ferguson. The theme of protest is also emphasized since protesters, protestors, protests, and protest are all mentioned. Information about crowds, riot, and pd is slightly deemphasized since these words do not have much frequency. 

My construction of a world cloud is relevant to the issues raised in the article “Big Data: Methodological Challenges and Approached for Sociological Analysis”by Ramine Tintati, Susan Halford, Leslie Carr, and Catherine Pope. According to this article, the few most influential people overshadow ordinary users, and therefore it is difficult for the people to make their voices heard (Tinati et al., 2014). In my results, the overriding voice was influential news reporters, instead of everyday activists. The article also discusses small-scale content analysis with Big data, which is what my results are a product of (Tinati, et al., 2014). My sample used tweets from a limited number of journalists and activists, but the number of tweets posted by these users was significant. Small-scale analysis made the data more manageable and easier to analyze, and observe trends and frequency.

 

Word Cloud