Tweet this, not that - Controversy Analysis on Twitter

Motivation

Social media has become a powerful tool for individuals and organizations to express their opinions and engage in discussions about various topics. Twitter, in particular, has emerged as a platform where users can share their thoughts on a wide range of subjects, from politics and social issues to entertainment and sports. However, not all topics are created equal, and some tend to generate more controversy than others. In this project, we set out to explore this phenomenon by analyzing controversial versus non-controversial topics on Twitter. Specifically, we focused on the levels of engagement that these topics receive, with the goal of understanding the factors that contribute to their popularity and impact. In this blog post, we will share our findings and insights from this project, shedding light on the complex dynamics of online discourse and how they shape public opinion.

Research Question: "Understand the patterns in controversial Vs Non-controversial topics"




Methodology

When it comes to studying controversial topics on Twitter, there are many challenges to consider. At the heart of our project was a desire to understand how people engage with sensitive issues on social media and what we could do to help mitigate the negative impact of these conversations. Here's a breakdown of our methodology:

Methodology

  1. Understanding the Problem Statement:
    To start, we researched existing literature on studying controversial topics on Twitter. This gave us a framework for understanding the scope of the issue and helped us identify areas of focus for our project. We then identified 2 types of user and make relevant personas by analyzing 50 twitter profiles selected randomly.

  2. Gathering Data:
    From our secondary research, we compiled a list of six topics - three controversial (abortion, gun control, and privacy) and three non-controversial (seatbelt use, child education, and sun exposure). We gathered data related to each topic by collecting 50,000 tweets on random dates.

  3. Doing Preliminary Analysis:
    With our dataset in hand, we analyzed how interactions varied among the topics and tried to identify interesting patterns. We looked for trends in engagement levels, sentiment, and language used in tweets related to each topic.

  4. Making the Model:
    Next, we created a model that could detect if a tweet was controversial or not and predict how much engagement it would get. We used regression and classification techniques to build our model.

  5. Refining the Model:
    While our initial model showed promise, we realized that it had limitations because we had only used six keywords. To improve the model, we trained it on a larger dataset for predicting likes and used OpenAI's text-davinci-003 model to predict if a tweet was controversial or not.

  6. Building an Application:
    Once our models were ready, we built an application using Flask. The application would detect if a tweet was controversial or not and then would rewrite the tweet to be less controversial. This application was created keeping in mind the personas created earlier.

  7. Offline Experiment:
    To see if our application would have an actual impact, we recruited two regular Twitter users and had them use our application. We then gathered insights on how our application impacted their social media experiences.

Results & Findings

From the initial analysis our findings were the following

In the world of social media, controversial tweets are bound to attract attention. However, this attention often comes with a cost. A recent study suggests that controversial tweets tend to generate more negative sentiment as compared to non-controversial ones.

Our analysis found that a majority of tweets in both controversial and non-controversial topics were actually neutral. However, when it came to controversial tweets, there was a significant increase in the number of negative sentiments expressed by users.

This finding sheds light on the impact of controversy in online discussions. While it may spark engagement, it also has the potential to generate negative reactions from users. As social media continues to shape our digital landscape, it's important to consider the impact of online discourse and how we can create a more positive and productive online environment

Sentiment Distribution of the tweets collected

Social media is an ever-evolving world, and it's no surprise that controversial tweets tend to generate more engagement than non-controversial ones. We found that in controversial tweets, negative sentiment tweets tend to get the most likes, views, and engagements.

In contrast, when it comes to non-controversial topics, neutral tweets are the ones that tend to receive the most engagement. This finding suggests that users engage differently with tweets depending on the nature of the topic being discussed.

It's essential to understand these patterns in social media engagement as they have implications for how content creators, marketers, and even policymakers approach social media. By understanding what types of tweets generate more engagement, we can create content that resonates better with our audiences, and ultimately foster a more productive online environment.

Engagement of Tweets

In both controversial and non-controversial topics, tweets with negative sentiment tended to come from accounts with fewer followers.

In the case of controversial topics, all negative sentiment tweets were found to come from accounts with less than 1000 followers. This suggests that users with a smaller following may be more likely to express negative sentiment in their tweets.
Who Tweets what?


We then made an application to help identify controversial tweets and predict the engagement. Our final model had an accuracy of about ~82%. The application is hosted here.

Entering the tweet


Analysis results


Rephrasing of Tweets

We then decided to conduct an offline experiment to gauge how much of an impact our application would have on actual twitter users.

Details of the Offline experiment

We noticed that users who used our app saw an increase in likes on their tweets. By rephrasing tweets in a more neutral and respectful manner, users were able to engage with their audience in a more positive way, leading to an increase in likes.

However, we also observed a decrease in the number of views for these tweets. This could be due to the fact that our app encourages users to tone down controversial language, which may make tweets less attention-grabbing.

Despite these changes in engagement, we found that there was no change in the frequency of tweets

How Likes and Views Changed from Rephrased Tweets


During our testing phase, we received some fascinating insights from our users. Firstly, users reported that the meaning of their tweets changed significantly after using our extension. By helping users rephrase their tweets in a more neutral and respectful manner, our app enables users to express their opinions without causing offense or misunderstanding.

Secondly, our users reported feeling more self-aware while tweeting. By providing suggestions for alternative phrasing, our app encourages users to think more carefully about their choice of words and the potential impact of their tweets on others.

Conclusion and Future Scope


The world of social media is ever-evolving, and it's important that we understand the dynamics of online interactions to create a healthy and productive online environment. Through our application that helps users rewrite controversial tweets, we gained valuable insights into the impact of online discourse.

When it comes to tweeting, it's important to consider the potential consequences of your words - especially if you're discussing controversial topics. While it's true that controversial tweets can generate a lot of engagement, there's also a risk of receiving negative feedback from those who disagree with your stance.

In fact, it's often those with fewer followers who are most likely to respond negatively to controversial tweets. This may be because they're less informed or less aware of the nuances of the issue at hand.

On the other hand, tweeting about non-controversial topics may not generate as much engagement, but it can have a positive impact on your personal or brand image. By sticking to neutral or positive topics, you can build a reputation as a trustworthy and respectable social media user.

Ultimately, the choice of what to tweet is up to you. Just keep in mind that there are potential benefits and drawbacks to both controversial and non-controversial topics, and it's worth considering these factors before hitting "send" on your next tweet.

As for the future scope of this project, we intend to create a chrome extension for our application as it is a lot more convenient to use. We also plan to improve our models by adding more tweets and getting better insights. Lastly we would like to thank PK sir and all the mentors we has in our CSS course for guiding us and helping us with the project.

-Pratyush Sharma (2022204010)
-Atharva Tendulkar (2022204008)
Karthik Konar (2022204012)
-Pravince Kumar (2022204023)
-Vivek Mathur (2020113002)
Our Team


Comments

Post a Comment

Popular posts from this blog

Parlia-metrics!

Parlia-metrics! Team Computational Anti-Social Scientists The Indian Lok Sabha debates are a rich source of political discourse, reflecting the complex and evolving nature of Indian democracy. These debates, which occur during the sessions of the Lok Sabha, offer a window into the key issues and concerns of Indian politics, as well as the rhetorical strategies and linguistic patterns of various political actors. In this project, we have leveraged the power of computational social science to analyze and model the proceedings of the Lok Sabha debates.  Goal The primary goal of this project is to delve into the vast realm of public-domain Indian data and unlock valuable insights from the transcripts of Lok Sabha debates. While datasets such as the Hindi Legal Documents Corpus (HLDC) and Indian Legal Documents Corpus (ILDC) have been explored and published, the parliamentary debates, despite being public, remain largely untapped in terms of comprehensive analysis and exploration. To be...
Read more

[READ THIS BEFORE YOU PITCH] The Indian Bible for The Indian Investor

Image
You have a great business idea, and you want to take it to the next level with further investment and support from people already established in the field. The logical next step would be pitching the idea and getting capital and support. What now? Pitching your startup to investors can be a nerve-wracking experience, especially if you're not familiar with the process. Even the best product can fail if it's not pitched properly and to the right person, which is why we decided to analyze patterns and trends from Shark Tank India data to gain insights into what factors influence a successful pitch. Our research question was simple: What factors contribute to a successful pitch? We approached this problem with a combination of linguistic and statistical methods. To start our analysis, we gathered datasets from Kaggle that contained metadata for seasons 1 and 2 of Shark Tank India. For those not familiar, Shark Tank India is a reality show wherein prospective pitchers or business st...
Read more

Stock Tip Simulator

Image
Introduction Though India still has only 3% of its population actively participating in the stock market, the total number of demat accounts has almost doubled over the past 10 years from 2.3 crores in 2010 to 5.3 crores in 2021. Most of these new investors are amateurs, and are looking to make quick profits. Such investors often resort to following tips from social media platforms like Reddit, Telegram and Whatsapp. On the other hand, the standard advice for amateur investors is to simply invest in index funds, since normal trades without enough background knowledge regarding the market and the company would often lead to losses. Through this project, we hope to understand which narrative is true and provide a tool that enables investors that choose to follow tips on social media to make informed choices.  We narrow down our scope to Telegram channels that give day trading tips. There are several public channels on Telegram with followers upwards of 1 lakh subscribers and we chose...
Read more