A Facebook civic hackathon has inspired open data leaders in Seattle to plan for new machine learning apps and algorithms for 2018.
In collaboration with city government, Facebook invited teams of technologists from local tech companies like Amazon to try their hand at civic tech. The Facebook hackathon pointed the coders at the city’s open data platform, and pitted teams against each other with the goal of solving the city’s toughest challenges with machine learning.
David Doyle, Seattle’s open data program manager, said the apps that resulted have convinced the coordinators that machine learning must play an influential role in the city’s open data program.
“As we enter 2018, we are encouraging all departments to think about how to potentially leverage opportunities to apply machine learning to their current and future open datasets based on the two use cases that emerged from the hackathon,” Doyle told StateScoop.
The two projects that won the hackathon were Find ‘n Park, an app that uses machine learning to help motorists find parking, and Contractor 5, software that assists people to find building contractors and cost estimates for new construction or remodeling.
Seattle’s permitting open data fuels Contractor 5 with project descriptions and estimates that developers say are within $5,000 of exact costs.
At the October event, the two teams showcased what could be done with open data as civic tech and smart city application evolve, Doyle said.
“It provided amazing insights into the power of machine learning and opportunities to think about a new generation of apps and services that could be powered by our open data,” Doyle said.
How and where the city will apply machine learning will likely be decided in the coming months in Seattle’s 2018 Open Data Plan, Doyle told StateScoop.
Looking at the larger picture, Doyle said the city also intends to begin considering machine learning as a broader piece of its operations.
“We are still thinking through how do we as a city provide the infrastructure — in terms of internal data science skills, cloud-based infrastructure and opportunities to partner with external groups — that can support such projects into 2018 and beyond,” Doyle said.