CURENT is collecting tweets containing climate change and electric power keywords as part of an ongoing project to analyze the impact of a power outage caused by a natural disaster on social media platforms. As of writing, there are almost 1B tweets with power keywords, and almost 45M on climate change. The problem is that most of these tweets are just noise; very few of the tweets collected with power keywords actually involve the power grid.
We need a good method for separating the relevant tweets from the irrelevant ones. Furthermore, we need to analyze the tweets that are relevant in the context of the this project (e.g. What is the sentiment? Is it in response to a natural disaster?)