HCII Associate Professor Jennifer Mankoff recently appeared on Pittsburgh's NPR affiliate, 90.5 WESA-FM, to discuss EDigs, a new CMU resource that uses machine learning technologies to estimate utility costs and help people find the perfect rental property.
"Prior to signing a lease, tenants are typically looking at location, parking, laundry — some of the basic needs that they have. But post-lease-signing, sometimes there are unexpected costs like unexpectedly high utility bills, or unexpected issues with safety, cleanliness, upkeep. All sorts of things come up once you start to live in a home that might not be as visible when you're first seeking a home," she said on the station's Essential Pittsburgh program.
"We're using data from the Residential Energy Consumption Survey, which is a survey that's run every couple of years by the U.S. government, and we're looking at the kinds of things that are predictive of utility costs in that survey, and then applying machine learning," Mankoff said. "Given the data that's in a housing advertisement, we can come up with a very rough estimate of your expected utility costs. Then, if a perspective tenant gives us more information about their habits and the home they are considering, we can improve those estimates further."