The 926th Wing’s Maj. Kimberly Hale, 14th Test Squadron, PM engineer analyst, won the second ever virtual Datathon, April 1, hosted by the Department of the Air Force Chief Data Office.
The Datathon challenge involved cleaning up large amounts of data from Aviation Resource Management in order to use Artificial Intelligence to help pilots optimize scheduling with the Puckboard aircrew scheduling tool.
The challenge was a collaboration amongst the Secretary of the Air Force Chief Data Office, AFWERX, DAF-Massachusetts Institute of Technology AI Accelerator, Lincoln Labs and the TRON software factory.
Participants spent five weeks working on the challenge in their free time, as well as using Slack as a shared workspace to share problems and solutions.
While 12 teams registered for the Datathon, only five made it all the way to the end to present their solutions and thought processes to a panel of judges.
Hale heard about the Datathon through the Air Force analytic community newsletter, as well as an AFWERX promotional email, who sponsors the planning and logistics of the program.
This problem in particular was chosen from a list of potential problems identified by DAF-MIT AI Accelerator, and given to AFWERX during one of their hosted workshops to choose the focus for this Datathon.
Hale said she was drawn to this challenge because she had seen the struggles first hand with scheduling on Puckboards when she walked into the 422nd Test and Evaluation Squadron during her last active duty assignment.
“I think because I had that little bit of history with Puckboards, I was like ‘Oh, yeah, this is a big problem’,” she said.
Hale felt prepared to take on the challenge coming from an extensive background of data science and analytics, to include her job in the AF Reserves and AD. She also completed a PhD program in the same realm that led to her becoming a data scientist in her civilian job.
“It was kind of right up my alley,” she said. “When I got out of Active Duty Air Force, I became a data scientist, that’s what I do in my civilian job.”
Although there was no monetary prize for this Datathon challenge, it did provide Hale with the opportunity to hone in on a particular programming skill, Python.
“It was a good use case for me, I had a timeline that was going to force me to finish something and practice with some intention,” she said. “That was one of my goals out of it.”
When asked if Hale would do more challenges in the future, she said she plans on it.
“If it’s something that aligns with my skillset, and I have time at the time of the challenge, then I usually try to at least investigate,” she said.
For now, the results of this Datathon are already being felt by the AF.
“The results of this Datathon are already feeding directly into Puckboard development and have a tangible impact on aircrew schedule and predictability and quality of life,” said Eileen Vidrine, Department of Air Force chief data officer.
TRON and DAF-MIT AI Accelerator developers will not copy the coding exactly as Hale has presented it, but will rather look at her thought process and solution to guide their development of the programing.
Vidrine pushed for the start of Datathon challenges in 2019 in the hopes that they could show the ability of Airmen and Guardians, both military and civilian, to help solve air and space problems in-house versus outsourcing when armed with the right tools and training.
“Challenges like this help our talented digital natives showcase their skills and explore new opportunities while bringing real value to the Department of the Air Force,” she said.
If you have problem that you believe can be solved with Datathon, reach out to Maj. Ryan Sullivan, DAF, Chief Data Office director of staff and Datathon lead planner for submission.
If you are interested in future challenges, make sure to look out for the AFWERX promotional emails, portal announcements, and Senior Air Force Leadership articles and social media channels.