Social Media and Policing

Social Media and Policing

Traditionally, Police all over the world have utilised a one-way communication model, sending information to the public either directly or through news media and not receiving communications back. Social media tools are changing these communication models, creating possibilities for interpersonal, participatory, and interactive communications.

Our project focuses on the use of the social media tool, Twitter, for the job of policing.We analysed the official Police handles of Mumbai, Bangalore, Delhi and Hyderabad on Twitter. The purpose of our analysis was to determine what type of information is shared by city police departments over Twitter and how the public uses the information shared to converse with the police departments and with each other.

Data Collection

We analysed 24,110 posts authored by the 4 city police departments and 2,31,589 posts of Twitter users who tagged these handles. The analysis showed that city police departments primarily use Twitter to disseminate crime and incident related information. City police departments also use Twitter to share information about their departments, events, traffic, safety awareness, and crime prevention.

Analysis

The analysis of posts was carried out using various tools. We looked at the most liked tweet, the most retweeted tweet, and found out that these are generally not the kind of tweets carrying some major information. Rather, the public tends to like a tweet more if it's a general in nature, with some witty text, or media content.


We also studied about the response times of police, and how long did they generally took to reply to the tweets of public that they were tagged in, and what kind of tweets got a faster response by the police.

We carried out Sentiment Analysis on the tweets posted by the public to study their nature, if they were positive, negative or neutral, to get to know the general opinion and consensus of the users. We also carried out a content analysis of the Tweets of the general public, to know what topics concern them the most.

We extracted the most commonly used hashtags and drew Word Clouds and Word Trees of the most relevant words to see how they are being used by the citizens. We also did an Age-Gender prediction on the tweets to find out more about our audience, however these are not the most accurate measures. 82.3% of the users were Male, and remaining Female, and 67.9% of the public lied in the age group of 25-34.

Results

We came up with the idea of introducing a Strategy Section, that discusses successful strategies that make an account popular among the users.

What is the proportion of tweets that should contain media content, what is the ideal response time that the police handles must restrain themselves to, what is the most active time of the day, etc. are some of the things we mention thereby.

We also had a Live Graph section, wherein we displayed the live graphs of the Follower Counts, the response times, the number of Tweets plotted against the time of the day, etc.

These graphs are helpful in getting real time statistics about all the handles.