Why are YOU responsible for George Floyd’s murder & Delhi Communal Riots!!

Original article was published on Artificial Intelligence on Medium

Why are YOU responsible for George Floyd’s murder & Delhi Communal Riots!!

A ML enthusiast’s approach to change the world.

Photo by munshots on Unsplash

“Social media is not a safe space.” — Tarana Burke

February and May were declared darks months for two of the biggest cities in the world; New Delhi and Miami respectively.

The article intends to show why YOU and ME are responsible for such events and most probably such events will happen in the future more often.

A Brief Introduction

Regarding American Riots

On May 25, 2020, George Floyd, a 46-year-old black man, died in Minneapolis, Minnesota, after Derek Chauvin, a white cop, bowed on his neck for very nearly nine minutes while he was bound to face down on the road. Two different officials further limited Floyd and a fourth official kept spectators from interceding. During the last three minutes, Floyd was unmoving and had no heartbeat. Officials did not endeavor to resuscitate him, and Chauvin’s knee stayed on his neck even as crisis clinical professionals endeavored to treat him. Two post-mortem examinations decided the way of Floyd’s demise to be manslaughter. (Source: Wikipedia)

Regarding Riots in Delhi

The 2020 Delhi mobs, or North East Delhi riots, were different floods of slaughter, property pulverization, and revolting in North East Delhi, starting from 23 February and caused predominantly by Hindu hordes assaulting Muslims. Of the 53 individuals executed, 66% were Muslims who were shot, sliced with rehashed blows, or set ablaze. The dead likewise incorporated a cop, an insight official, and over twelve Hindus, who were shot or attacked. Over seven days after the brutality had finished, many injured were moping in deficiently staffed clinical offices and carcasses were being found in open channels. By mid-March, numerous Muslims had stayed missing. (Source: Wikipedia)

Some Psychological facts about Society

  • The presence of other people can have a powerful impact on behavior. At the point when various individuals witness something, for example, a mishap, the more individuals that are available the more outlandish it is that somebody will step forward to help. This is known as the bystander effect.
  • People will go to great lengths to obey an authority figure. Individuals will go to incredible, and now and then hazardous, lengths to obey authority figures. In his famous obedience experiments, psychologist Stanley Milgram found that people would be willing to deliver a potentially fatal electrical shock to another person when ordered to by the experimenters.
  • The need to conform leads people to go along with the group. A great many people will oblige the gathering, regardless of whether they think the gathering isn’t right. In Solomon Asch’s conformity experiments, people were asked to judge which was the longest of three lines. When other members of the group picked the wrong line, participants were more likely to choose the same line.
  • Sometimes it is easier to just go along with the crowd than cause a scene. In gatherings, individuals frequently oblige the larger part conclusion as opposed to cause disturbance. This marvel is known as groupthink and tends to occur more frequently when bunch of individuals share a lot in like manner when the gathering is under pressure, or within the sight of a magnetic pioneer.

Project Blueprint

Before having the option to spread out an outline, a compact goal is required.

The goal is, I will be creating a Machine Learning model that will fetch all tweets related with both incidents for as far back as three days, and will perform sentiment analysis just as detest discourse acknowledgment (on similar information), to assess popular feeling concerning Americans and Indians for respective countries.

For Riots in America

  1. Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “Racism”.
  2. Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “George Floyd”.
  3. Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “White People”.
  4. Fetch all tweets that have been posted from May 26th to May 28th, for the keyword “Black People”.
  5. According to that score, perform sentiment analysis as a means of dictating public opinion for Americans.
  6. Perform Hate-Speech Recognition and Analysis of similar tweets and direct the harsh level of racial maligning.

For Riots in Delhi

  1. Fetch all tweets that have been posted from Feb 24th to Feb 26th, for the keyword “Islam”.
  2. Fetch all tweets that have been posted from Feb 24th to Feb 26th, for the keyword “Hindu”.
  3. Fetch all tweets that have been posted from Feb 24th to Feb 26th, for the keyword “NRC-CAA”.
  4. Fetch all tweets that have been posted from Feb 24th to Feb 26th, for the keyword “Delhi-riots”.
  5. According to that score, perform sentiment analysis as a means of dictating public opinion for Indians.
  6. Perform Hate-Speech Recognition and Analysis of similar tweets and direct the harsh level of religious defamation.

For purpose of tweet fetching for required keyword and required time, we will be using Tweep(official API for twitter), for that one need a consumer key and access token that can be easily availed once you turn on developer’s option.

The searchTerm denotes the required keyword, NoOfTerms denotes total tweets we wish to fetch.

To fetch tweets, we use tweepy.Cursor(), it is an iterator item and result can be yield using iterator function.

As a result, I have successfully fetched 10065 tweets related to “George Floyd”.

It is now important to assign a sentiment score to each individual tweet. This can be performed with the following code:

Using textblob library, we can analyze every tweet based on its polarity and subjectivity.

And based on the aggregate score, we will classify the tweets in Negative, Neutral and Positive classes.

To understand data in a more clear and better way, we can plot a “WordCloud” that gives us a fair idea about frequent words used.

The word cloud created for American riots is :-

The word cloud generated for Delhi riots is :-