Many of us grew up watching Star Trek. Whether it was the crew of the original Enterprise or the Next Generation, we watched the officers on the bridge remotely probe a planet’s surface for clues about its society and level of development. We’ve already seen many Federation realities come about – communicators, hypospray, and tricorder-like devices. And, now, technology can provide valuable insights about buildings and properties in much the same way the Enterprise crew gained information about the structures on the planet below.
Recognizing the Difference Between Catastrophic and Evolved Damage
Storylines in the Star Trek franchise often started with “remote scans” of a planet. Here on Earth, as resolution and frequency of observation data improve, we find that remote scans really do tell a story. As the resolution of imagery has improved, we’ve gained the ability to “see” critical property details, such as roof or surface conditions, construction methods and materials, and existence and quality of vegetation. As the frequency of observations has improved, we have greater understanding of how events shaping the property have unfolded.
This information is particularly valuable for insurance companies to determine whether damage occurred due to a catastrophic event or if it evolved over time. For instance, using a grading method of good/fair/poor, we track roof condition. Thus, if a homeowner complains of a leak following a storm, an insurer can counter that it was likely leaking before the storm because of its poor condition.
Similarly, insurers can better manage their portfolios by understanding a property’s risk factors. We identify the existence of features like swimming pools or trampolines, and also quantify hard-to-measure features such as the amount of tree-overhang that encroaches structures on the property.
Precision Estimating … From Your Desktop
Improved resolution and the availability of imagery in major metro areas give us the ability to automate tasks, such as exterior inspections and many types of estimating.
A byproduct of our detailed land classification is exact measurement of all property attributes, including turf, impervious surface, water features and structures. This information, combined with proprietary measurement and identification attributes, allows geoanalytic companies to automate estimating tasks such as roof repair/replacement, landscaping bids, and road maintenance.
So far, most of these advances have used ortho or nadir images – imagery that is taken from above. The aerial imagery industry essentially grew from the need for high resolution “satellite” imagery. However, new imagery companies are emerging that look at properties from all angles. This will give companies the ability to measure other building features like siding and windows.
Unprecedented Insight Into Markets
Most marketing research is based on demographics, such as economic status, gender, or geographic location, and incorporating demographics into marketing plans has dramatically increased return on investment for advertisers. However, geoanalytics can provide an even finer granularity and further improve ROI.
For instance, Arlington and Fairfax counties in the DC suburbs have nearly identical socio-economic demographics. Yet, residents in Fairfax are nearly four times as likely to have a yard as their neighbors in Arlington. So, if you’re in the landscaping business, you would not want to waste your advertising dollars in Arlington.
Geoanalytics help companies in numerous industries better target their marketing outreach by identifying concentrations of features like swimming pools, solar panels, or yards. When layered on top of traditional demographics, geoanalytics becomes a valuable marketing tool.
Fascinating … But How Did the Enterprise Crew Compile a Planet’s Worth of Data?
Applying “analytics” to remotely sensed imagery is really just doing the same work you were already doing, but with additional detail … until you add machine learning into the mix.
OmniEarth does more than just apply analytics to remotely-sensed imagery. Using the same type of machine learning techniques that NASA and the Government (and, presumably, the Federation) use, our data scientists train computers to do the work of an analyst – teaching them to classify property attributes and perform measurements. This not only ensures consistency across measurements, it gives us the ability to rapidly identify property attributes – as much as 40 times faster than could be done by a human analyst.
This artificial intelligence, when combined with cloud computing services like Amazon Cloud Services or IBM Watson, enables us to scale our products across large regions or even across the entire country. Consequently, it’s not that far of a stretch to believe that we will be able to count buildings of a particular size, style or material across the entire planet in the not-so-distant future.