Climate Change & Global Sentiment
How do environmental changes impact global wellbeing as measured on social media?
What is the impact of climate events, such as temperature increases and environmental disasters, on subjective wellbeing and happiness? This project aims at addressing this issue by creating a comprehensive Global Sentiment Index (GSI) based on social media data.
The MIT Sustainable Urbanization Lab has collected years of geolocated social media posts from Twitter and Weibo. These entries are encoded into vectorial representations using several Natural Language Processing (NLP) techniques, including dictionary-based methods (LIWC, Emoji dictionaries) and embedding-based algorithms. Pre-trained machine learning algorithms use these representations to impute the Twitter data’s sentiment and main topics, every day and everywhere.
We intend to combine these social media time-series with environmental data in order to answer a series of questions.
Research Question 1:
What is the effect of climate change and environmental disasters on subjective psychology?
We intend to consider the link between the effects of climate change and human psychology. Using social media, we can estimate the sentiment change due to:
- warming temperatures.
- increased weather unpredictability.
- more frequent environmental disasters.
Matching historical data on the climate metrics to the GSI, and controlling for seasonality and news cycles, we can estimate the effect of climate change on subjective psychology.
Research Question 2:
How has belief in Climate Change evolved over time?
We also intend to study how climate change is discussed on social media: in particular, when do people acknowledge the reality of climate change and the importance to do something about it? Among the phenomena we can analyze are:
- when climate change is discussed seriously vs. jokingly.
- which subtopics are associated with serious climate change-related conversations?
- based on sentiment and key terms, how impactful (in terms of likes or retweets) is climate change-related content.
Using topic-modeling techniques, we can restrict the universe of Twitter data to climate-related content. LIWC will enable us to impute sentiment on these tweets (with more nuance than positive/negative). The general time series of sentiment will give us an evolution of the way Climate Change is discussed on social media. We can also evaluate the most effective communication on the subject by comparing sentiment and reach to the key terms used in the tweet.