Machine Learning Shows YOU Are Responsible For George Floyd’s Murder!
The practical 9 minutes eye-opener needed to bring change once and for all
“ I really have no words. My sole intention all this time was to spread awareness and combat racism. Little did I know, I was contributing to it. ”
As the majority of you know by now, May 25th marked the murder of 46 years old Minnesota resident George Floyd. What sets this incident apart, is the actors involved in his death, as well as the incentives that appear to have been the fuel for this attack. This 6 minutes article is meant to open your eyes and show you how you practically share responsibility for this depictable act.
The 46 years old man was murdered in one of the most inhumane ways I have personally ever seen by a Minnesota Police officer. Although there is some background preceding the incident, it is irrelevant to the topic of this article. What is of importance, is that the man posed no threat to anyone, and after having committed a “harmless” crime, suffocated from the pressure induced by the officer’s knee on his neck.
The crime has been categorized as a hate crime. The officer was not acting on behalf of the police force, but rather on behalf of his own personal insecurities and unexplainable bigotry.
What is of importance though, is that YOU, the person reading this, makes it possible for such an event to not only take place but also most probably repeat itself in the near future.
As a Data Scientist and Psychology enthusiast, I felt it was my duty to show you exactly what you are doing wrong and how to fix it, using Science and Data.
A Quick Walkthrough on The Psychological Aspect Of The Research (it is not that long, do not worry)
While performing the Machine Learning analysis, it will become apparent that my conclusion will be based on significant psychological research. The aforementioned research is based on a scientific paper called “Neural and Behavioral Correlates of Sacred Values and Vulnerability to Violent Extremism. Frontiers in Psychology.” and can be accessed from here.
This is a thorough paper, so instead of explaining the entire background, I will only be sharing the important parts.
The goal was by using fMRI scans, to explore underlying neural and behavioral relationships between sacred values, violent extremism, and social exclusion. Ethnographic fieldwork and psychological surveys were carried out among 535 young men from a European Muslim community in neighborhoods in and around Barcelona, Spain. Candidates for an fMRI experiment were selected from those who expressed willingness to engage in or facilitate, violence associated with jihadist causes; 38 of whom agreed to be scanned.
Experiment #1 - Cyberball
The subjects were initially asked to play a game called cyber-ball ( a game known for yielding strong feelings of social exclusion). After a couple of rounds, when players started to feel excluded, they were placed in an MRI scanner.
The aim of the experiment was to observe through the MRI machine, how the participants’ brains reacted after being exposed to social exclusion.
It became apparent that when the subjects were exposed to such phenomena, there was increased activity at a certain part of the brain. This same part is usually activated when a person has radical thoughts.
Conclusion of the First Experiment:
When excluded, the list of things the subjects were willing to fight/die for grew bigger.
This only refers to the people who had already been at the blink of radicalization. The effect on people who have had no prior exposure to such notions is dramatically worse. Hate is induced in their systems and it rapidly brews until it becomes deadly. People with no predisposition towards hatred for a specific group suddenly start to shift their views completely. In the end, if enough pressure is induced, this hate reaches levels unprecedented to the individual. The result is radicalization.
Experiment #2 - Peer Pressure
The aim of the second experiment was to test how the subjects reacted to peer pressure. The rules of the experiment were the following:
- 30 Participants were presented with a scale from 1–7
- They were asked to rate their willingness to fight/die for certain values.
- They were then presented with where their peers were placed on the scale while being observed by an MRI scanner and were asked to re-consider their responses.
Below there is an image of the subjects’ brains before and after being told where their peers ranked on the scale:
Conclusion of the Second Experiment:
When asked to pick a value on the scale, that part of the brain was completely idle, showing a lack of willingness to negotiate/be persuaded.
When on the other hand they were told that the wider social group was placed on a lower value than them, that part of the brain was re-activated and they changed their answer to match that of their peers.
*This part of the brain is known to be associated with a person's openness to new ideas and persuasion.
What can we assume by observing the results of the two experiments?
Continue reading to understand why you were and will be responsible for future similar acts of violence and bigotry!
Project Blueprint
Before being able to lay out a blueprint, a concise objective is needed.
The objective is clear, I will be building a Machine Learning model that will scrap all related-to-the-incident tweets for the past three days, and will perform sentiment analysis as well as hate-speech recognition (on the same data), in order to evaluate public opinion concerning people of the white race.
The steps I will be following are the following:
- Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “Fuck White”.
- Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “George Floyd”.
- Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “White People”.
- Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “Whites”.
- According to that score, perform sentiment analysis as a means of dictating public opinion for white people.
- Perform Hate-Speech Recognition and Analysis of the same tweets and dictate the rough percentage of racial defamation.
Performing The Analysis
The initial components of the model (fetching and cleaning the data e.t.c.) are the exact same as the ones I have used here (you can follow the instructions step by step, and no alteration of the code is required):
As a result, I have successfully fetched 109,641 tweets related to the keywords presented above.
It is now important to assign a sentiment score to each individual tweet. This can be performed with the following code:
By creating a “WordCloud” we can see some of the most common and significant words among the tweets.
Cross Referencing with Racial Hate Speech Model
It is crucial to have a model with accurate results. In order to correctly categorize the sentiment scores, I will be cross-referencing the results with a Hate Speech Detection model.
The dataset used, as well as some informative literature that helped me significantly, can be found below:
Results
The results of the model are as alarming as I anticipated.
Out of the 109,641 different individual tweets, more than 44% were racially targeting Caucasian individuals and using racial slurs against them. Another 27.8% was sympathetic without using any type of racial slurs and the remaining 27.5% of the tweets were neutral and did not mention anything about different races.
To better comprehend the results, the total numbers for each category are:
Negative 49,017
Positive 30,488
Neutral 30,136Why is this important?
It all comes back to the title of this article. “Why am I most probably accountable for this attack and the attacks to follow?”, you may ask.
According to the experiments we explored earlier, it became clear that blaming entire groups and effectively segregating people can only have negative effects. In reality, by racially swearing at white people, where the majority of them are completely against what happened that day, can make them more susceptible to violence. This phenomenon was clearly showcased with the first experiment of cyber ball.
In essence, if there were individuals who for any reason were sympathetic to the belief system of the murderer, going on social media and reacting the way the majority of people are currently reacting, would only increase his/her willingness of perpetrating an act of terror and hatred.
“ Can we stop this? “
The answer to this question can be provided by the second experiment. Instead of trying to marginalize white people and call them accountable for the tragedy that took George’s life, white people must by themselves state in masses that they are against such incidents of racially motivated violence. The reality is, that the overwhelming majority, (the 99% of the 99% of the 99%) does not condone the actions of that Law Officer under any circumstances, and most probably find themselves being disgusted by the mere thought of it.
By uniting ourselves and mass condemning the actions OF THE PERSON who killed George, as well as the belief system he holds dear, can we bring real change. Only then, will these individuals placing themselves on “Seven”, come down to “Two”.
I hope this article has provided you with a vital lesson on how we must move forward as the human race. It is in our power to stop such phenomena from becoming the norm. The only thing required is to act collectively as one first and not choose to see the things that separate us, but rather the things that unite us.
Do you want to learn more?
If you want to advance your knowledge and are interested in making money using Machine Learning I highly encourage you to follow me and read the articles listed below:
