Three top contenders edged out more than 3,800 teams from 91 counties to qualify for the Second Round of a Zillow Prize® competition and the chance to win a $1 million grand prize if they succeed in building an algorithm to predict actual sales prices. Winners will be announced in early 2019.
The three top teams from Round One are among 100 teams who were challenged to make incremental improvements in Zillow’s Zestimate. The top contenders include a three-person team, each from a different country, who had never met each other, a father-daughter duo from Switzerland, and a solo data scientist from Japan. Collectively, they were awarded $50,000 in prize money.
Since announcing Zillow Prize in May 2017, the competition has become one of the most popular machine learning competitions on Kaggle, a platform for date-related competitions and the world’s largest community of data scientists.
To qualify for the Second Round, teams had to improve the accuracy of the Zestimate beyond the current error rate of 4.2 percent in the three competition counties (Los Angeles, Orange and Ventura, all in California).
Stan Humphries, creator of the Zestimate and Zillow Group’s chief analytics officer said his company hopes the prize inspires current data scientists to dive into the world of real estate valuations and inspires a future generation to consider a career in data science and machine learning. “There’s an incredible opportunity to bring new and diverse minds into the field,” he remarked. Regarding the data science community’s response to the Zillow Prize, Humphries said there were lots of innovative solutions. “We’ve been blown away,” he commented.
Zillow is matching the first-round prize money with a $50,000 donation to #YesWeCode, a Dream Corps initiative. #YesWeCode helps 100,000 young women and men from underrepresented backgrounds find success in the tech sector. This is part of Zillow Group’s greater efforts to support STEM-focused organizations which includes donations and investments in the University of Washington School Of Computer Science and Engineering, DiscoverU, and the Washington State Opportunity Scholarship.
Zillow publishes Zestimate home valuations on more than 100 million homes across the country based on 7.5 million statistical and machine learning models that examine hundreds of data points on each individual home. Calculations for Zestimate use data from county and tax assessor records and direct feeds from hundreds of multiple listing services and brokerages, plus input from homeowners.