Success story
UC Davis and Max Planck Institute
Using Twitter data to study how people adapt to climate change.
Here's the tl;dr
How are humans adapting to extreme weather and climate change? Do they adapt without even realizing it? Frances Moore of University of California, Davis, and Nick Obradovich of the Max Planck Institute for Human Development partnered to evaluate over 2 billion Tweets sourced from the Twitter API’s streaming endpoints. Their paper explores how people have perceived increased extreme weather events driven by climate change.
2 billion+
Tweets analyzed for the study
Endpoints
Challenge
How are people adapting to extreme weather and climate change? Do they get used to it over time? How quickly do people normalize weather changes and does that play into how seriously they take climate change? These were the driving questions that Frances Moore, Assistant Professor in the Department of Environmental Science and Policy at University of California, Davis, and Nick Obradovich, Senior Research Scientist and Principal Investigator at the Max Planck Institute for Human Development, set out to ask.
For years, climate scientists and social scientists have been wondering if humans are undergoing the “boiling frog” effect: referencing an apocryphal scenario where if you hypothetically put a frog in a pot of water and slowly bring it to a boil, the frog will not notice as the temperature is rising. Are humans simply adapting to climatic changes without even realizing it? The challenge in answering these questions was large. How could one gather a large enough dataset in real-time from the public, and avoid introducing bias as people - if asked directly - may not be as forthcoming when interacting with a scientist?
Moore and Obradovich partnered up to solve this data challenge, and quickly realized that the Twitter API could be the tool they were looking for to answer the question when it comes to how people’s expectations and adaptation relates to climate change.
Solution
Before connecting with Moore, Obradovich had been using the streaming endpoints on the Twitter API to collect data over the years to understand how people’s sentiment changed over time when exposed to different weather. “It was very serendipitous [when Moore reached out with her research question], we had already created classifiers for parsing weather-related Tweets from non-weather related Tweets, and had worked on understanding the sentiment of these Tweets,” said Obradovich. “We had validated these data in varying settings, even across different social platforms. So when [Moore] reached out, we realized we could map our approach differently over to our dataset of Tweets to answer the research questions [Moore] had been pondering.”
Together, Moore and Obradovich spent two years evaluating over 2 billion Tweets that were geolocated in the continental United States, pulling out Tweets related to weather. They measured peoples’ sentiment as well as the ‘remarkability’ of different temperatures (the frequency of talking about weather) and how it changes with repeated exposure to unusual temperatures.
In addition to the mapping these classifiers over the Tweet dataset to perform sentiment analysis, the team went on to merge in climate/weather data, and aggregate Twitter data to different spatial levels – such as by county or major cities that have seen extreme weather events.
Leveraging Twitter data sourced from streaming endpoints on the Twitter API, they were able to answer the boiling frog question and published their findings, which have been widely received.
"We’re trying to get at some of the negative effects of climate change that are widespread but not necessarily disastrous...Twitter can give us this aggregated measure of what those social consequences are."
Francis Moore, Assistant Professor in the Department of Environmental Science and Policy at University of California, Davis, CA
Results
The study found that people use the last two to eight years as their reference point for “normal” weather — and that remarkability of unusual weather declines with repeated exposure. What people consider “normal” weather quickly changes based on what they’ve experienced in the last few years — even though that weather may have significantly changed over the course of their lifetime. As people adapt to a constantly evolving “new normal”, they may not be cognizant of how climate change is creating those changes. And worse, they may not feel the need to act.
“We did not answer the question about whether a decline in remarkability impacts the likelihood that people will take action about climate change,” said Moore. “But the findings of our study do show that people very quickly normalize changes in weather and keep normalizing it, even when those changes may not be ‘normal’ at all.”
But the research made them more curious. “This adaptation to extreme weather, or a normalization effect that occurs, also raised questions about the lasting impact of this extreme weather on your physiological and mental health,” said Obradovich. “We found that - though people adapted their discussions of the weather to repeated changes - there was little evidence that the effect of hot temperatures on sentiment adapted in a similar way. In other words, people were still affected by weather extremes, even as they ceased remarking on them.”
Looking forward
Moore and Obradovich have gone on to use Twitter data to evaluate the extent and impact of coastal flooding, using information from Twitter to supplement more standard environmental monitoring instruments like tide gauges. “We’re trying to get at some of the negative effects of climate change that are widespread but not necessarily disastrous. They’re having these negative consequences across large populations, they’re interrupting people’s daily lives, they’re annoying, and they’re causing some damage,” said Moore, as quoted here. “Maybe not a huge amount, but those types of impacts are pretty important as people are trying to go about their daily lives.”
Moore goes on to say: “The main reason I like [Twitter] as a source of data is that it integrates not just a measure of typical exposure — which is ‘Did the water come onto the land in a place where it wasn’t supposed to be?’... [But also] measures ‘What are people noticing? What are people talking about?’ Twitter can give us this aggregated measure of what those social consequences of that particular flood are.”
The scientists plan to expand their research. “While the bag of words classifiers worked fine in this climate change case, we want to do more. We are looking to improve our use of NLP (natural language processing) technologies to better identify types of speech, and there are very good deep learning models. Such models are the future of where we’ll go to figure out what people are talking about and how they’re talking about it,” said Obradovich.
Interested in learning more? See their research publications:
Explore more
See how others have used the Twitter API.