As the US Conference of Mayors convenes this week in Washington DC, we (Open Data Nation) are taking a bold bet that the smart cities of tomorrow will need predictive statistics, including machine learning, in their tool belt.
Reactive is not good enough, if you can be proactive.
There are billions of records of city, administrative data, and the pace of this data generation is accelerating exponentially as new IoT sensors are installed and with connected and autonomous vehicles about to take the road.
It’s not good enough anymore to describe the problem, when there are enough data to reliably predict and prevent problems from happening in the first place. For example, in Vision Zero cities, we should no longer rely on analytics that show X people died in a car crash last year, and black and brown people are 7x more likely to be struck . We can and should put our data to work (what we call “purpose-driven data”) to anticipate crashes and save lives.
You can read more about how Open Data Nation is working with Microsoft to bring street safety to the next frontier of smart cities HERE.
Do the math
Statistics is kind of scary. Few people look fondly back on their mandatory high-school experience. But the power of statistics is transformational. With it, we can achieve what we set out to when we published data to begin with:
- Plan — Know where the problem is going to be tomorrow, not yesterday
- Prioritize — Right-size, and pinpoint exactly where and when (in real-time) resources should be spent to optimize for impact, rather than spreading it thin or randomly
- Evaluate — Controlling for all the dynamism of cities, know whether what you did worked and whether the money you spent worked to have an impact
It’s a smart-city evolution (not a revolution)
Not every city is currently ready for machine learning. We know this because we asked cities across the US, from Anchorage, Alaska to Miami, Florida, where are you today, and where do you need to be tomorrow?
- Initiative: Developing an action plan with community engagement
- Descriptive: Analyzing and visualizing the magnitude of a problem
- Active: Responding with activity in known problem areas
- Adaptive: Planning ahead and learning how and where to optimize for impact
If you need help along the way, funny thing, we’re on a mission to bring machine learning to your city’s tool box.
This blog post originally appeared on Medium.