Tracking Tweets with FME

The TweetSearcher in FME runs a search for Twitter entries. By entering a coordinate You can also request geocoded tweets within a specified searchradius.

As the Swedish Christmas Goat succumbs to flames (once again) I came up with the idea to investigate that specific area – in Gavle Sweden. So I quite easily created a FME workbench that starts by sending the coordinate of the (former) Christmas Goat to the TweetSearcher.

I was quite surprised by the amount of actual geocoded Tweets that got returned. The TweetSearcher also returns user information and timestamps. By sorting this info and sending it to a PointConnector I could recreate Paths of the different users who have Tweeted in the area.

Each line in the below picture represents a user who have tweeted at least twice at different times and locations:

Tracking the Railway Tweets

An interesting pattern emerged along the railway. I decided to examine this a bit more. In the below picture I have masked somewhat of the usernames. If You enlarge it by clicking You can clearly see that there have been some tweeting along the tracks:

By playing up the actual Tweets along the track an interesting monolog appears:

Simple translation:”Come on train. I don’t want to spend christmas in Gavle”
Simple translation:”Weird that traincompanies have all that power…”
Simple translation: “Won’t get any further. Train decided that I should spend christmas here….”
Simple translation:”Come on train. I don’t want to spend christmas in Gavle”

The passenger seems a bit disapointed due to some delay in the train schedule. (Personally I would not mind spending christmas in Gavle). Without knowing it this have been very helpful to verify the precision of geocoded Tweets….

Leave a Reply

Your email address will not be published. Required fields are marked *

nineteen − six =