Objective: To understand geographically related tweets.
Methodology:
-
Aggregate all tweets per hashtag to the county geography in United States
-
Understand distibution of raw tweets per county
-
Normalize the raw counts with population of the county
-
Log transformation of the data to better suit geographically weighted regression methodolgy
-
GWR on each combination of the tweets to understand basic observations
- Basic observations:
- Map matrix (all hashtags vs. all other hashtags)
- Top twelve hashtags correlations weighted geographically for each tweet -