EconTwitter is one of my favorite communities on the internet---a place to find new papers as well as new recipes. So I thought it would be fun to create the EconTwitter network that represents this community, so that anyone can examine what it really looks like.
This is the universe of Twitter economists. Of course, the boundaries defining an "economist" are hard to define. I created this dataset from two sources:
- The RePEc list of economists on Twitter.
- This represents the largest official list of economists on Twitter that I could find
- Repeated #EconTwitter tweeters.
- I defined this as anyone who tweeted with the hashtag #EconTwitter more than twice in the 30-day period before September 14th, 2020 (when I created the dataset).
- This consists of both RePEc economists and people who are active on EconTwitter even if not officially listed (e.g. because they are predocs/PhD students/working in industry)
When building the EconTwitter network, this is the node list.
The variables in econs.csv are:
| Variable | Meaning |
|---|---|
| id | The user's numerical ID |
| name | The user's display name |
| handle | The user's @ handle |
| following | How many people does the user follow? |
| followers | How many people follow the user? |
| verified | Is the user verified? |
| favorites | How many tweets has the user favorited? |
| join_date | When did the user join Twitter? |
| object | A Python stringified object representing other user attributes |
| is_human | Is the user human, or a bot/institutional account? |
| econ_following | How many economists does the user follow? |
| econ_followers | How many economists follow the user? |
| following_ratio | What fraction of the user's follows are economists? |
| followers_ratio | What fraction of the user's followers are economists? |
| in_deg_centrality | What is the user's in-degree centrality in the network? |
| out_deg_centrality | What is the user's out-degree centrality in the network? |
| avg_followers_of_following | How many followers do the user's follows have on average? |
This is a pickled file, readable only in Python using NetworkX. It is a directed graph representing the EconTwitter network. I built this network iteratively from the econs.csv node list. The generation process is straightforward: search through the follow-list of every node, and create an edge from X to Y if Y is in both X's follow-list and in the node list.
This is an edge list for the EconTwitter network, created indirectly from econtwitter.gpickle. It simply saves every source-target pair in the network, so that the network can be recreated outside of Python.