Public Reaction towards Black Lives Matter Campaign
On May 25, a 46-years-old black American man named George Floyd was killed in Minneapolis, Minnesota, during an arrest for allegedly using a counterfeit bill. He was forced to the pavement and restrained by two police officers while a third white senior officer, Derek Chauvin, knelt on Floyd’s neck for nearly eight minutes. This incident triggered a big wave of protest against police violence toward black people across the United States. Not only in the US but the protest also happened in other countries. This mass movement then came to be known as the Black Lives Matter campaign.
In this article, I would like to share my finding regarding the Black Lives Matter campaign specifically from the public perspective. The public opinion itself then would be classified as one of three sentiment groups i.e positive, neutral, or negative. In addition, Named Entity Recognition (NER) analysis also has been used to give a better understanding of what happened in the Black Lives Matter campaign.
To do so, I used a supervised machine learning algorithm i.e logistic regression to analyze the public opinion that has been addressed in the Twitter platform from June 4 to June 26, 2020. The tweets have been crawled using Twitter streaming API and using keywords like “George Floyd”, “blacklivematter” and “racism”. I also make sure that the tweets are written in English. This condition is purposely added to make sure that the tweets are coming from the United States region. Python and SparkNLP are the main tools that I used in this research.
Sentiment Analysis
Let’s start with the sentiment’s share. As seen from below pie chart, most of the crawled tweets are written in a negative tone. Those tweets almost dominated all of the tweets crawled with a share percentage of about 73%. While for the positive and neutral the percentages is almost similar.
Since Black Lives Matter is a campaign that seeks justice for black people and has been triggered by the tragic murderer of George Floyd, it makes sense that there are a lot of people angry and express their anger on Twitter. Next, let’s see the movements of each sentiment category toward a period of time.
There is a significant increase in the number of tweets with a negative tone between June 13 to June 16. I did further research about what happened in that particular period of time, and I found that there is another tragedy occurred. A 27-year-old African American man named Rayshard Brooks was reported to be shot by Atlanta Police Department (APD) office Garret Rolfe on the night of June 12. I assume this event is the main reason why the public rage significantly increases.
Word Cloud Analysis
The above picture is the word cloud from all the tweets with a negative tone. Using this word cloud I can justify what kind of word that people frequently used in their tweets. There are harsh words like died, hung, and killed. I assume that this word mostly came from news’ tweets that then have been retweeted by many people. As far as I know, there is massive news coming after the incident of George Floyd. By retweet this kind of tweet, people might express their support toward the Black Lives Matter campaign.
While for the word cloud built using tweets with a positive tone, I found terms like police reform and justice act. Although there are still many words that look like harsh words, here in that word cloud still can be found positive words that might indicate people hope. The people hope and demand reformation in the police department.
Named Entity Recognition Analysis
Lastly, I conduct a Named Entity Recognition (NER) to find out which entity frequently occurred during the conversation of the Black Lives Matter campaign. This analysis itself will be more focused to find out either person or location entities only.
Below is the graph that shows the top 10 entity person from the collected tweets. Also, the word cloud was built using all entities that have been classified as persons.
Besides George Floyd and Rayshard Brook that have been mention earlier, I also found another name i.e Dominique Fells and Eric Garner. I am not that familiar with these two names so I did a little research about them. Dominique “Rem’mie” Fells, is a black transgender woman from Philadelphia whose body was found on the bank of the Schuylkill River with a stab wounds and trauma in the head and face. According to CBS Philadelphia, the authorities believe 36-years-old Akhenaton Jones is responsible for Fells’ murder. The other name is Eric Garner, who is also a victim of police violence toward black people. Eric died in the New York City borough of Staten Island after Daniel Pantaleo, a New York City Police Department (NYPD) officer put him in a chokehold while arresting him. This happened on July 17, 2014.
From that result, I draw a conclusion that the Black Lives Matter campaign is not only about George Floyd but also about all the black people who have been treated unfairly either by police officers or the communities. Not only for what happened recently but also for what happened in the past as well.
While in the top 10 entities locations, I found some region names such as Minneapolis, Houston, Washington, and Atlanta. According to Vox, the campaign itself has spread worldwide since the protest also can be found in London, Seoul, Sydney, Monrovia, Rio de Janeiro. As you can see, there are entity name UK and India in the top 10 entities.
This is the end of this article. Any findings and conclusions in this article were made by me and there is a lot of possibilities that it might strong with my subjectivities. In other articles, I will share how to set up Python, Twitter Streaming API, and Spark as the main tools to do this kind of research.
Thank you for reading!
