September 23 to 25, 2022
AI for Good is a 48-hour hackathon bringing together UNF student teams, and partners JEA, FIS, and Microsoft to build solutions that drive social and environmental benefits.
The hackathon was organized using Devpost. Follow the below Devpost link to view winners, judging criteria, rules, and schedule.AI4Good Devpost Site
Over the course of a weekend, UNF students collaborated to work on unique problems impacting the environment. Students were provided with the tools, support, and data needed so that they can write the code to right the world.
Participating students received mentorship from FIS, Microsoft, JEA, and UNF experts. Participants were expected to use their skills in the problem domain, programming, data mining, machine learning, statistical analysis, and visualization techniques to solve the problem. Hackathon participants were expected to form a team with a maximum of three to work on the hackathon problems and datasets.
JEA provided two problem areas and relevant datasets for student teams to work during the hackathon.
Predict solar farm power generation given historical data.
Optimize a vehicle routing problem given starting location, ending location, and a number of maintenance stops that must be made.
FIS was the lead organizer for the hackathon event handling logistics, food, and prizes.
Microsoft provided access to Azure services and in-house support to use Azure technologies to solve the hackathon problems. Microsoft covered costs for prizes.
Friday, September 23
Saturday, September 24
Full day of coding and hacking remainder of the day and through the night
Sunday, September 25
Predicting Future Solar Power Output from NOAA Weather Data
by Jeremy Cantor, Aidan Doruska, and Jack Swingle
Solar prediction: Predict solar farm production from weather
by Marion Forrest, Troy Kidd, and Jason Swinehart
PF: Your optimal pathfinder
by Michael Kimollo, Prajwol Lamichhane, and Kris Roker
Dr. Lakshmi Goel
Each participating team was evaluated on their model performance and presentation of the results. Winners of the hackathon was determined based on the below evaluation weightings and ratings.
Model Performance (40%)
Solution Approach (40%)
5 = Excellent presentation on all counts
4 = Very good presentation with minor issues
3 = Pitch can be significantly improved but is satisfactory overall
2 = Pitch is incomplete and severely lacking
1 = Presenter fails to address the criteria
Halley Den Nys
Dr. Xudong Liu
Dr. Indika Kahanda
Dr. Michelle DeDeo
Dr. Richard Shang
Matthias van Meurs