Imagine having millions of citizens sharing real-time data on damages and hazards into one view.
​​
During a disaster, citizens can use their mobile devices to upload photos, GPS coordinates, and detailed reports of damages, blocked roads, and hazards. This crowdsourced information can seamlessly integrate into an ArcGIS map, giving utilities a comprehensive view of the entire city to enhance decision-making in their operations centers.
​
With machine learning models trained to predict restoration times and automated job assignments, we can transform the way utilities operate at scale during a disaster. By combining AI-driven job prioritization, rapid data collection, and real-time updates, we empower cities and their utilities to accelerate recovery efforts while providing citizens with crucial information about available services during prolonged outages.
About
I live in Houston and was one of the >2 million people who lost power to Hurricane Beryl. Thankfully, the online community helped me find critical information like where to find gas or a hot meal, but it was clear there was a better way to find this information without scrolling through tons of social media threads.
I was initially inspired to mock up a mobile app concept to make it easier for communities to share information after a disaster, but credit goes to my husband for observing we could also send in damage reports to our utility. Building on this further, I realized there were a lot of information gaps preventing a utility from being able to restore power quickly to an area hit with a disaster.
If I were in the utility operations team, and I only knew which of my service areas were out of power, how would I know what to do next? How much damage was done to the infrastructure servicing that area, to tell me how many service men to send to the site? If there were fallen trees affecting the damage site, do I have the vegetation crew en route to clear it so that linemen can repair it? Are the roads blocked by debris or flood waters, or hazards like fallen lines, and where are they located? If there are teams arriving from out of town, how do I communicate and coordinate across all of them at scale? How do I prioritize what gets worked on, especially considering the need to restore critical services like hospitals, police and fire departments?
​
I don't have experience with what goes on behind the scenes of restoring power, but I'd hazard a guess that all of the things I've listed are relevant to any utility trying to operate under crisis.
​
While I still know very little about how a utility operates in these scenarios, I still believe there is a lot of potential to using crowdsourced data - we could have over 6M citizens providing details compared to using 600 official inspectors who would have to traverse disaster conditions to safely complete damage assessment. With cell service (mobile towers were powered by backup batteries during outages), citizens can upload photos, coordinates, and details on damages, blocked roads, and hazards near their location across the entire city. This information can be integrated into an ArcGIS map so that utilities can have an overlay that feeds into their operations center.
​
Machine learning models can then help to automate job prioritization based on historic restoration times, utility set rules (e.g. preference to restore critical services), and find opportunities for quick wins where there are easy fixes and available service teams nearby. It could relieve the burden on utility operations teams trying to handle a sudden surge in scale, complexity, and pressure in the aftermath of a disaster. Providing citizens with an easy way to send information, and offering updates through a single platform, can help relieve some of the stress of going without power and set expectations.
​
Of course, none of this works without the utility, integrations into their systems, and a deep understanding of their workflow. If I were a founder attempting to build this, the first challenge is getting the right connections. The next would be checking against existential risks that can kill this as a business - does the utility already have something like this (either in-house or 3rd party that's already ready to go)? Will each utility need a custom built solution (which would drive up costs and erode margins)? What kind of budget do they have?
​
My main hope is that we won't see another Beryl-level outage soon.