Two teams from Cal Poly placed first and second in competitions at DataFest, a statistics hackathon at UCLA.
One team, consisting of statistics junior Andrew Voorhees, statistics junior Brian Bahmanyar and mathematics junior Zach Zhang, took first place in the “best insight” category. The other team, consisting of statistics sophomore Jonathan Kisch, business administration junior Shreyas Doshi and statistics sophomore Jonah Muresan took second place in the “best use of external data” category. Almost 20 teams from five colleges, including UCLA and USC, took part in three separate competitions.
“This showcases what our students can do,” statistics professor and the team’s faculty advisor Hunter Glanz said. “That’s only going to boost the image of statistics and other related departments at Cal Poly.”
At the beginning of the weekend, the teams were presented with data from Ticketmaster, including web traffic data and ticket sales, and had less than 48 hours to present their findings, Glanz said.
“There were millions of records and many different variables, and it was up to students to dig and investigate the data at the event and try to find something interesting that Ticketmaster didn’t already know about their demographic, sales, audience or data,” Glanz said.
The team that won the “best insight” category made predictions between artists and locations to help Ticketmaster predict ticket sales more accurately in given locations for certain artists.
The team that took second place for the best use of external data created an interactive map that identified where Ticketmaster is underpricing or overpricing tickets based on per capita income by county.
“This would only increase sales, because if you have location premium you are able to charge more money without losing demand necessarily, and if you have a location discount you are making the event more affordable to the locality there,” Kisch said.
According to Glanz, DataFest is a unique opportunity for students to get more experience outside of the classroom.
“We don’t usually expose students in their coursework here to datasets that are that big, real or rich. There were millions and millions of observations of records in these data sets,” Glanz said. “They also aren’t usually given a data set and told ‘find something interesting.’ They have to choose what to use, what to apply and how to work on things on their own.”