3/25/2021

Problem Description

  • Politicians’ public image is important, and particularly so if they are elected officials.
  • How can politicians represent themselves and their ideas or opinions to the public?
  • Twitter is a powerful communication tool with voters, and one that is increasingly relevant to politics.
  • Using Twitter to gauge voters’ attitudes towards politicians could be useful.
  • Proposal: In this project, we will be measuring the affective response (and precursors) to certain tweets by politicians and relating those to shifts in their approval ratings over time.

Data

  • We will be gathering data using Twitter’s API.
  • Focusing on tweets by presidents and the replies.
  • Relate this data to various polling data regarding the president’s approval ratings.

Analytics Plan

  • Analytics

    1. Building sentiment profiles for the public reactions.
    2. Machine learning models to predict changes in approval rating based on changes in the affective responses to tweets.

Evaluation Plan

  • Qualitative

    1. Visualizations on relative aggregate sentiments of responses to each tweet.
    2. visualize the shifts in various polls estimations of presidents’ approval ratings.

  • Quantitative

    1. Building profiles for the affective response to tweets.
    2. Modeling the relationship between response profiles and shifts in approval rating.
    3. Prediction of shifts in approval rate based on affective response.
    4. Check Accuracy (RMSE R^2..etc).