@Officer1070 is a robot highway patrol officer that simulates immigration status checks with its Twitter followers based on Arizona traffic stop data.
Arizona Daily Star staff compiled the dataset for a series investigating enforcement of SB 1070, the Arizona law that requires law enforcement to verify the immigration status of someone they’ve detained if they suspect the person is in the country illegally.
During a Migrahack event last year, I worked with Star reporters Luis Carrasco and Perla Trevizo and University of Arizona student Amanda Martinez to build No Timely Response, an interactive that puts users in the role of the highway patrol officer during an immigration check. @Officer1070 turns the tables, placing the user in the position of the driver.
The bot is an attempt to explore SB 1070 and the traffic stop dataset in a different way. From a technology perspective, I had wanted to make a Twitter bot for years but struggled to come up with an idea that didn’t spam the Twittersphere.
@Officer1070 was inspired by two particular projects:
- Darius Kazemi’s bots – especially @sortingbot, which places its followers in Hogwarts houses
- NPR Waterbug tool, which gave me the idea for generating images
More specifically, tweepy’s streamer constantly listens for a follow event and sends follower data to a function that creates the tweet and image.
Pillow creates the image by writing text over a background image with the pixel coordinates I defined for each text box.
I probably struggled the most making the script run constantly on a remote server, but Chris Keller alerted me to the nohup command, and it seems to be working so far.
I realize 2016 may be rather late to be launching my first Twitter bot. Some Twitter bots might be more appropriate performing tasks in dedicated chat rooms. Kazemi himself said changes to the platform, like a rumored update to character limits, could derail artistic bots.