Bob Bennett, the CIO for Kansas City, Mo., felt confident that a series of sensor initiatives was helping him and his team reach their goal of creating “the smartest city on Planet Earth.” Then he realized “while that was really cool, it wasn’t smart because it didn’t take advantage of the collective wisdom inside City Hall and it didn’t take advantage, quite frankly, of the data that we already had resident in City Hall.”
So the city, which collects more than 4,500 datasets daily, started working with Xaqt’s MetroGraph platform. MetroGraph integrates data from multiple sources, including open, historical and sensor data, in a single quantitative analysis environment. Translation: The platform ingests the information, and Xaqt data scientists write algorithms on top of it to produce visualizations that update in real time.
On June 23, Xaqt and the city are rolling out their latest application: anomaly detection. As the platform ingests the data, a baseline of what is normal forms, Xaqt founder and CEO Chris Crosby explained. “We sort of create a fingerprint, if you will, for each aspect of the city,” he said. “As that data’s flowing into our system, we’re looking for variances outside of the norm. Those get thrown into our anomaly dashboard.”
Using that dashboard, city workers can see in real time if there is a spike in 311 calls regarding missed trash collection in a neighborhood, for example.
In the coming weeks, anomaly detection for crime, traffic and parking will be released. This means city workers will be able to see that a traffic incident has occurred before anyone calls the police because they’ll be able to note a change in traffic flow, for instance. What’s more, users can subscribe to get alerts via Cisco Spark or Slack messaging services and automatically create chat sessions around the data.
“This revolutionizes all city services because now we can predict a problem that will be newsworthy six hours from now,” Bennett said. “The ultimate result of this, we believe, is going to be a city that anticipates the needs of its citizens, based both on historical trends and what the data is showing us. It’s going to be a city that manages resources based on what those needs are going to be so we can be better stewards of the public funding that we do receive.”
Another algorithm helps Kansas City predict where potholes are likely to form based on data such as historical weather data, 311 reports of potholes and road condition data. It shows that potholes typically develop 77 days after ice melts and the water penetrates the asphalt.
“If I know that I’ve got several areas in a particular road that are likely to be damaged in about 77 days, at about Day 30 I can run this other truck that has a little radar on it to that section of road and I can say to myself after sending that radar out, ‘Yes, indeed, this will be a problem in the near-term future because I can look in the road and see the health of it,’” Bennett said. That leads to significant cost savings because the city has time to order a 55-gallon drum of asphalt for $20 rather than a 5-gallon bucket of quick-curing goop for several hundred dollars.
“I’m saving money there and I’m scheduling it so the citizens who live on that street are not going to be negatively impacted by an unexpected road closure,” Bennett said. “Chris is literally revolutionizing the way we manage streets.”
Besides using a city’s existing or streaming data, the platform taps into data from national sources such as the Census Bureau and the Securities and Exchange Commission, Crosby said. To access data that isn’t open, such as from sensors, Xaqt has partnered with companies such as Cisco Systems, GE and Verizon to build supported integrations.
Article originally published here.