Dive into the details on our novel approach to tree canopy data

December 27, 2021  | Alec Sabatini

Our new subscription-based, artificial intelligence-driven tree canopy data is disrupting the traditional tree canopy mapping and assessment process that has been the status quo for over a decade. Naturally, that is generating some discussion around methods, accuracy, human intervention, and data ownership.

We’re glad to be a central figure in the conversation and at the leading edge of the industry. We’d like to further that discussion by exploring some of the recurring topics we are seeing come up around Urban Tree Canopy and Artificial Intelligence.

Personal Attention Where It Matters Most

Our mission is to ensure every community has access to high quality, affordable tree canopy data and planning tools. At the core of this is our novel approach of using artificial intelligence (AI) and machine learning to develop high-resolution land cover data. Understandably, some have assumed that means a reduction in accuracy, with less hands-on involvement from GIS experts. Fortunately, that’s not the case. AI does not replace the human element of canopy analysis. It refocuses our team’s time and attention where it is needed most.

Analyzing tree canopy data used to require weeks of meticulous quality control by GIS technicians to reach a sufficient level of accuracy. Our new, nationally-available canopy data has the same level of accuracy or better, but is ready almost instantly. That means our GIS experts spend less time on data quality and assurance checks that can be subjective and potentially inconsistent and more time on data interpretation and analysis. The end product is a more insightful assessment that is ready faster, updated more frequently, more consistent from one year to the next, and costs less.

For example, we recently completed a tree canopy assessment for 32 Florida cities that were in the path of Hurricane Michael. Using traditional methods, the project budget would have only supported a basic change analysis. With the time and cost savings of AI-derived canopy data, our Geospatial Mapping team was able to do more analysis for more cities and build a detailed, interactive story map to share the results of their work.

Our data accuracy is still best in class, with a proven accuracy level of 96.6% or above using 48,000 points of random sampling.

Top Tier Accuracy, Just Like Before

Over the last few years, there has been an increase in companies offering tree canopy data, including some major tech players. Now that there are more sources for tree canopy data, accuracy is becoming a key differentiator between providers.

Though our methods are evolving, our data accuracy is still best in class, with a proven accuracy level of 96.6% or above using 48,000 points of random sampling. By using advanced machine learning algorithms instead of manual labor, more common errors are avoided.

Our subscription model makes it affordable for communities to complete a tree canopy analysis every 2-3 years, the frequency that new statewide aerial imagery becomes available, instead of the conventional 5-10 years. By using the exact same modern analysis methodologies year after year, there is also greater reliability when comparing canopy change from multiple time periods.

We are happy to show you the canopy data in your city so you can see for yourself. It’s part of our process. Please get in touch if you are interested in learning more.