The U.S. population is racially diverse- a characteristic necessary for racial injustice. So, I expect that the majority of #BlackLivesMatter tweets to be from 30N-40N and 90W-150W — the United States. I am interested to see if #BlackLivesMatter tweets were pulled from my home town at 40N 95W (if I were to guess our coordinates).
To date, 83935 tweets were pulled from Twitter.
I have 131 mappable tweets, which means that 0.16% of the tweets have geolocations. Nathan Yau, in his book Data Points, writes about the importance of telling the story that we intend to tell with data. Therefore, it is necessary for me to mention that the map of #BlackLivesMatter tweets (below) is by no means an accurate depiction of support of the Black Lives Matter movement throughout the world. Rather, the map below is an investigation of mappable data and data visualization. From the geolocations that are available, 30W-40N 70W-90W, I expect most of the tweets to be from the U.S. Moreover, all but about 10 tweets are written in English. I expected most of the tweets to be in English, because to my knowledge most black people in the U.S. speak English.
Of all of the hashtags being monitored in class, #BlackLivesMatter, with roughly 83,000 tweets, ranked second in most number of tweets. This would suggest that the set of #BlackLivesMatter tweets is big data. However, Kate Crawford, in Think Again: Big Data, writes that “there are “signal problems” in big-data sets — dark zones or shadows where some citizens and communities are overlooked and underrepresented.” So it is possible that some of the people who would be of interest to study, people of color in poor urban areas, may not be represented in my data set. This is of particular concern since #BlackLivesMatter is heavily concerned with people of color in poor urban areas. Crawford also writes that “only 16 percent of online adults in the United States use Twitted, and they are by no means a representative sample — they skew younger and more urban than the general population.” This fact suggests that the “dark zones or shadows” of big data may not affect as severely a study of young urban people. It is obvious that the data collected needs to be studied further before it can be concluded that the data is “good data”.
The other hashtags on my team are #noDAPL and #woke. To date, #noDAPL has 27,000 tweets and #woke has 11,500. Perhaps #BlackLivesMatter has 4-8 times the number of tweets, because injustices covered by #BlackLivesMatter are long standing. The Black Lives Matter movement began in 2013, and continues to be a addressed in the news and on social media. On the other hand, #woke is more of a trend than a movement. As for the Dakota Pipeline, my guess would be that #MuslimBan has overshadowed #noDAPL.
The map above features the 131 mappable tweets of the total 83935 tweets pulled from Twitter starting February 1 and ending February 15. Each yellow circular marker (size 1) represents a single tweet. As expected, most of the mappable tweets were from the United States. I was surprised that there were a few tweets from Germany, UK, South Africa, and Malaysia. Since only 0.16% of the tweets are mappable, I expect that many more than just the mapped tweets are from people outside of the U.S. #BlackLivesMatter tweets also appear to not exists in the mid-west of the U.S. This may be the result of many factors: limited access the internet, less support for the Black Lives Matter movement, smaller non-white population, or these people have turned off Twitter’s access to their location. It is frustrating that we cannot know more about why there are so few mappable tweets. Making conclusions about the data has proven more difficult than I expected. This map could be telling a story about the support of the Black Lives Matter movement across the world, or it could be telling the story of where people are less sensitive to whether their location is shared by Twitter.
It appears in the map above that there is a concentration of tweets from San Francisco, CA 94102 and Brooklyn, NY 11217. Therefore, the races of the people living in these areas in 1970, 1990 and 2010 were mapped below.
I expected the mapped tweets to have been posted be people of color. Consequently, I expected the population living in San Francisco, CA 94102 and Brooklyn, NY 11217 to be predominantly people of color.
As seen in Figure 7 and Figure 8, the most predominant race of people living in San Francisco, CA 94102 and Brooklyn, NY 11217 during all three years was white. The white population has decreased with time, but still the black population was and is fewer than the white population. This may suggest that white people are also using #BlackLivesMatter. However, it seems unreasonable to make an assumption on the majority race of people using #BlackLivesMatter based on 0.16 mappable data.
Black Lives Matter Allies in Change, written by Alex Tom, Margi Clarke, Preeti Shekar, Karina Muñiz, Megan Swoboda, Felicia Gustin, and Devonté Jackson, captures the opinions and action taken by groups in support of the Black Lives Matter Movement. These groups include the Chinese Progressive Association, Asians 4 Black Lives, Latinas 4 Black Lives, Bay Area Solidarity Action Team, Showing Up For Racial Justice, Black Alliance for Just Immigration. Many of these groups mention the importance of black people in the U.S. in the fight for racial and immigrant justice. It seems logical that people of color would be the most prominent supporters of the Black Lives Matter movement. Therefore, it still makes little sense to my why the mappable tweets come from prominently white areas.
Generations of Struggle, written by Percy Green II, Robin D. G. Kelley, Tef Poe, George Lipsitz, Jamala Rogers, and Elizabeth Hinton, sheds light on the commonly perceived gap between Martin Luther King Jr. and the Black Lives Matter Movement. There were in fact other movements and institutions such as the Black United Front, the Organization for Black Struggle, and the Black Radical Congress. I wish that it were possible to do a data study of tweets from the 1970’s. This is of course impossible, since Twitter did not exists in 1970. However, as Generations of Struggle suggests, we may find that there are more similarities than differences in how people protested for black justice in 1970 and today.
Software-sorted geographies are places in which algorithms allocate social or geographical access to goods, services, and opportunities for mobility (Graham). Usually, only when it is profitable are are goods, services and opportunities allocated to people. Consequently, groups of people, who already do not have access are prevented from ever gaining access. The internet is no exception. Therefore, I will need to be careful not to perpetuate the methodology behind the algorithms that create software-sorted geographies, when I study the Twitter #BlackLivesMatter data.
Graham, Stephen D. N. 2005. “Software-Sorted Geographies.” In The People, Place and Space Reader, eds. Gieseking, Mangold, Katz, Low, Saegert, 133-138. New York: Routledge.
Green, Percy, Robin D. G. Kelley, Tef Poe, George Lipsitz, Jamala Rogers, and Elizabeth Hinton. “Generations of Struggle.” Transition, no. 119 (2016): 9-16. doi:10.2979/transition.119.1.03.
Tom, Alex, Margi Clarke, Preeti Shekar, Karina Muñiz, Megan Swoboda, Felicia Gustin, and Devonté Jackson. “Black Lives Matter Allies in Change.” Race, Poverty & the Environment 20, no. 2 (2015): 26-32. http://www.jstor.org.ezproxy.trincoll.edu/stable/43873217.
Yau, Nathan. 2013 . “Data Points: Visualization That Means Something.” 61-68.