Geospatial analytics is a huge and growing market. It’s been estimated that the global geospatial analysis market grew to $12 billion by 2020, with an annual growth rate of 16%. This blog post will teach you the basics of geospatial analytics, specifically for property, and it’s uses across business settings. You’ll learn what this field entails and how companies can use it for their own purposes.
To start, geospatial analysis is the study of analyzing data that is geographically referenced. This type of data includes things like GPS coordinates, location information, and proximity to other features. The word “geospatial” is derived from two words, geo and spatial. The term “geo” means earth or land, while the term “spatial” refers to a location in space. Geospatial analysis is an analytical technique that involves gathering data about the Earth’s surface (such as elevation, population density) or the built environment (roads, buildings) and using them to make predictions about human behavior or natural phenomena. ESRI is one of the best-known companies in the ‘GIS’ space. Generating geospatial analytics takes this a step further, with the goal of finding patterns and meaning within geospatial data, that can help improve decision-making, business performance, and power operational efficiencies. The creation of modern analytics necessitates computing power, which makes the field well-suited to AI.
So what is geospatial analytics?
Geospatial analytics is a form of computational analysis that leverages geographic information, spatial data, location data, and increasingly, high-resolution imagery, computer vision, and other forms of AI to extract structured data that can be used for specific applications and industries.
Geospatial Analytics: What It Is, Why It Matters
The roots of geospatial analytics can be traced as far back as the earliest forms of cartography and surveying, if not much further. But beginning in the 1960s, the term became associated with geographic information systems (GIS) used to catalog natural resources. Since then, geospatial data has been used in a wide variety of ways, and can span many forms of technology—satellite imagery, GPS, coordinate systems, location sensors, and additional location-based data.
Governments use geospatial data for everything from tracking extreme weather events, to measuring and managing traffic patterns, to assessing population growth and its impact on energy, transportation, and housing needs as part of urban planning. And the military has long used geospatial data to assess resource placement.
More recently, however, the revolution in cloud computing and advances in machine learning have enabled an ever-growing amount of geospatial information to be analyzed at high speeds—helping to power decision-making and whole new business processes.
As part of this revolution, geospatial imagery has become an important source of truth on the grounds that it makes geospatial analysis easier to verify and explain. As the quality and availability of geospatial imagery has increased, so has its utility in a new generation of geospatial analytics-based solutions. Many of these solutions are building a better understanding of a central aspect of the global economy: property.
Property data has always been challenging to collect, and for property-related industries such as insurance and real estate, the impact of geospatial analytics on organizations is nothing short of transformational. For example, instead of the time- and labor-intensive task of conducting onsite inspections, critically important information on exterior property condition is increasingly available on-demand.
We’ll go deeper into all of this in just a bit. But first, let’s look at how geospatial property analytics is created, and why it’s beneficial in key use cases.
How Geospatial Analytics Works
Powered by AI, modern geospatial analytics use high-resolution geospatial imagery and computer vision to extract structured data from imagery. When focused on property, this includes the size of a property, type of roof construction and its condition, the presence of solar panels, whether there are overhanging trees, and whether there’s a backyard swimming pool, yard debris, or vegetation encroachment around the property that could pose a fire danger, and more.
Today’s most advanced solutions can provide this kind of data for tens of millions or even a hundred million commercial and residential properties nationally, delivered with the speed and coverage of traditional property record data.
By providing instant, accurate intel on property conditions, modern analytics enables users to assess a specific property, as well as compare it to similar properties within the same neighborhood, zip code, city, region, or by type across the nation. They can also analyze local, regional, and national trends that can impact its value or risk profile over time.
The value of this intelligence depends on the resolution and recency of the imagery it uses and the machine learning put in place to analyze it. To understand why, let’s start by examining different types of geospatial imagery used for various kinds of analysis. As we explained in a recent post, different image-capture methods come with distinct pros and cons. And manual review of these images by analysts presents its own issues.
Remote Sensing Data & Geospatial Imagery
Remote sensing is a technique that uses imaging from satellites, aircraft, and other sensors to capture information about the Earth. This technology has been used for decades to gather data on natural resources such as water quality and forest cover. It also provides critical inputs into many fields like meteorology, agriculture, disaster management, and military intelligence. Remote sensing data is one of the key inputs into modern geospatial analytics, alongside other geo-referenced data.
There are different forms of remote sensing data—but geospatial imagery is the most powerful for property-related analytics. Geospatial imagery is available at different resolutions and refresh rates. Today, common forms include satellite imagery and aerial imagery captured through manned aircraft.
Unmanned aircraft such as drones can provide very high-resolution imagery, but have limited coverage. Today, manned aircraft ranks among some of the best resources for high-resolution geospatial imagery with wide coverage of highly-populated areas every few months. So the images have excellent data currency at a reasonable price. Increasingly, stratospheric balloons are becoming another source of high-resolution imagery with recency extending to a few days, and perhaps someday, real-time.
Machine Learning & Computer Vision
What are the hidden patterns in your data? In a world of data, spatial data analysis is often the first step to understanding and analyzing geo-referenced data. However, it can be difficult for analysts to go about this process at scale as most tools are designed for smaller datasets or single users. Traditionally, statistical modeling methods such as regression techniques were used to make predictions and analyze data, but with the rise of machine learning and computer vision, we can use algorithms to create far more powerful and accurate analyses that can identify new insights and correlations we otherwise would not be able to achieve. By using machine learning and advanced data science techniques, geospatial analytics are further turned into predictive analytics, which can spot trends as they take shape.
Use Cases: From Imagery to Intel
Of course, “straightforward” is not the same as “easy.” Transforming imagery into structured data (i.e., useful information) is intricate and requires a substantive level of expertise.
At CAPE, for instance, we begin by knitting together a coast-to-coast view of the United States using geospatial imagery provided by our imagery partners. Then, we deploy tailored computer vision algorithms to identify key elements in the imagery relevant to risk analysis and valuation. Next, additional data, ranging from parcel information to recent weather events to hazard information, are layered in via machine learning to create a comprehensive view of a given property. Finally, the outputs are returned as structured data linked to an address.
Next, let’s look at some solutions to understand how this kind of AI-enabled geospatial analysis can be put to use—including within the insurance and real estate industries.
The business intelligence gleaned from geospatial analytics help insurers select better risks, reduce expenses, and improve the customer experience in a number of ways.
– Residential Property
According to McKinsey, the use of digital channels to buy and manage insurance advanced more than five years during the COVID-19 pandemic. More than 30% of consumers that have ever used a digital channel to buy or manage insurance did so for the first time after March 2020. The success of digital channels rely on a combination of quick and accurate quoting, binding, renewing, and claims management. On-demand access to key property attributes is a critical ingredient in all of these workflows.
Carriers like Kin are turning to geospatial analytics for the most recent and accurate property information, giving them the confidence to accurately price risk when offering online quotes to customers shopping for homeowners insurance online. Geospatial analytics also helps insurers make faster and more accurate underwriting decisions while reserving physical inspections for situations that require additional scrutiny.
It also helps insurers make faster and more accurate underwriting decisions while reserving physical inspection for situations that require additional scrutiny.
For example, homeowners policies on properties with a bad roof have had loss ratios three times higher than those on properties with roofs in good or excellent condition. In fact, 40% of insurance claims are related to the roof. Roof Condition Rating (RCR), created by the team here at CAPE, analyzes every roof on more than 110 million structures nationwide, rating them on a five-point scale ranging from severe to excellent. All available through API, and is easily integrated with underwriters’ existing workflows. It’s also worth pointing out that this is also without the need for expensive, time-consuming physical inspections.
Then there are renewals. After a property is underwritten, the profile on that property is often never updated again. By incorporating timely geospatial property data into renewal workflows, insurers like CSAA reduce inspections of roofs known to be high-quality, provide more accurate pricing at renewal, and mitigate future losses by helping members proactively repair damaged roofs. mitigate risk, like cutting vegetation back, before damage occurs. repair damaged roofs. With automated change detection, insurers can also know when the risk profile of a property has meaningfully changed and right-size insurance coverage to offer the best protection.
– Commercial Property
Consistent, structured property condition data enables commercial property insurers to power more scalable workflows. Solutions offering at-a-glance condition scoring and visual highlights of structures that need attention are especially valuable. Our own customers report that at least 20% of past underwriting or inspection decisions would have been different if more accurate property condition data had been available. And over a quarter of those policies, or 5% to 10% of the average book, would have been declined out right or denied renewal, because they did not meet existing eligibility criteria.
External property condition data unearthed through geospatial analytics is also correlated to overall property risk and internal claims. Building roof condition, for instance, is linked to higher wind and hail losses and business interruption claims. Property condition is linked to liability. And lot debris is linked to property damage and liability losses. In addition to helping insurers avoid unnecessary inspection costs, post-bind pricing changes, and unexpected, downstream losses, geospatial analysis is starting to enable innovative new underwriting strategies.
– Natural Catastrophes: Wildfire, Wind, and Hail
Whether it’s pure sprawl, the effects of climate change, or both, wind and wildfire are increasingly important factors in insurer risk evaluations. Homes and businesses with a large amount of foliage overhanging a roof experience 90% higher wind-related losses, for instance. And in the western US, a tremendous number of homes are being built in areas with high wildfire risks. While these settings are certainly bucolic, the mix of beauty and human encroachment make for a truly combustible mix. Indeed, wildfire is one of the most granular perils in terms of how risk can deviate from one property to the next.
Modern geospatial analytics solutions can provide insurers with on-demand access to structured data on important wildfire attributes down to the specific property level, including proximity to vegetation and surrounding fuel load. According to our own data, for instance, homes with heavy vegetation coverage have 115% higher wildfire claim frequency and 272% higher overall losses compared to homes with defensible space.
What’s more, homes in very high hazard zones with heavy vegetation coverage have 5X higher losses of those with more defensible surroundings. By identifying property attributes such as roof geometry, insurers are better able to capture premium, triage inspections, and avoid unexpected losses.
Just as with insurers, AI-driven geospatial analytics is rapidly gaining the attention of the real estate industry for real estate projects, precisely because viable alternatives fully understanding property risk are often not available, not granular enough, or are outdated. According to Deloitte, these technologies represent a quick, lean, and affordable way to gain an understanding of assets all the way down to zip code, neighborhood, and individual-address level. And this can dramatically enhance decision-making for interests throughout the real estate ecosystem, including the following.
– Loan Trading
Geospatial analysis is increasingly used to augment traditional data with additional condition variables that can significantly improve valuation and pricing models for better-informed decisions on loan investments, origination and underwriting, and more. For many, broker price opinions (BPOs) and appraisals are too costly and time-consuming for a marketplace that demands split-second decisions. And even sophisticated automated valuation models (AVMs) can miss critical details that impact financial decisions.
Using geospatial analytics, loan traders can identify outlier properties and quickly determine whether target properties are in good condition. Loan traders can also flag potentially risky properties and estimate potential rehab costs and identify properties that have incomplete information in order to help determine more accurate purchase prices.
Finally, loan traders can optimize or expedite the property inspection process using this information while monitoring existing positions to make sure collateral stays in good condition.
– Multi-Property Investment and Management
These and other factors are of particular interest to investors competing in what has become a red-hot market for single-family rentals (SFRs). Back when investing in SFRs was primarily a local affair, filling in the information gaps in traditional sources of property information might have been somewhat straightforward. But as new technologies enable larger real estate investors to easily manage numerous properties at once, bigger players and billions of dollars in capital are pouring into the sector.
Because traditional AVMs and even drive-bys miss critical and current property features that impact rehab cost and value, their usefulness in the new world of SFR investing is limited on their own. Whether they’re exploring a single property, or shopping citywide, regionally, or even nationally, geospatial analysis enables investors to instantly access neighborhood, census tract, or zip code, or CBSA-level data so they can zero in on the most promising regions and properties.
They can also review aggregate changes over time to unearth trends (both good and bad) that might otherwise be challenging to identify in order to decide which markets to enter.
– Loan Origination and Servicing
Originators and servicers can use geospatial analysis to better value and assess property. They can use geospatial intelligence to identify properties that differ from existing data, with new fields such as roof condition, pool, solar panel, and number of structures and auto-fill or crock-check property features and characteristics.
They can optimize or expedite the property inspection process and automatically understand if there are important changes to property condition that may affect the collateral value.
Endless Additional Possibilities
These are just a few examples of how we at CAPE apply geospatial analytics today. But there are many other potential use cases for these data-drive technological capabilities. Roofing, solar panel installation companies and others can use it to conduct remote inspections for prospecting and quoting. So can property and landscape management companies.
Local governments can conduct property tax assessments far more efficiently. Telecoms and other utility companies can identify towers at risk from vegetation overgrowth to manage clearance operations. And geospatial analytics’ value to FEMA and other emergency response organizations may be immeasurable in the aftermath of catastrophic events. For geospatial analytics, the sky’s the limit.
Geospatial Analytics is ‘Explainable AI’ at Its Best
According to Deloitte, full-scale “macro-to-micro” geospatial analytics is now mature enough to be adopted by enough users for it to make a real impact in 2021. This is in part because geospatial solutions can do more than just remove blind spots—they can dramatically improve the accuracy of predictive models and the evaluation of risk.
Since these technologies are based on images or other remote sensing technologies, they are in some ways the ultimate form of “explainable AI,” which is AI-applied in such a way that the intelligence it provides is easy to understand. In other words, geospatial property analytics can help power the right amount of insurance premium, or why one property is worth 15% more than a similar one nearby.
By “knowing what others don’t,” geospatial analytics is likely to deliver a considerable competitive advantage to those who harness it first.
To learn more about geospatial analytics and its use cases in industries such as insurance and real estate, visit capeanalytics.com