Understanding Development Trends through Updated Parcel Data

Understanding Development Trends through Updated Parcel Data

Combining Orbital Insight land use and open-source parcel data helps identify outdated parcel data or unregistered property

  • Governments require an up-to-date understanding of land use, development trends, and parcel data in their area. Maintaining key datasets such as road networks, buildings, and agricultural land can be expensive and time-consuming since methodology consistently differs from one locality or even department to the next. Accurate parcel data informs everything from zoning maps to tax assessment, timber management, and emergency operations.

Data from satellite imagery is a widely-used resource and often the foundation for existing datasets, updates and validation practices. However, many of these sources are often incomplete or outdated, particularly in fast-growing or rural regions. For local governments, there is a twofold challenge at work: (1) such imagery can be subject to infrequent or inconsistent collection across different regions, and (2) substantial time is required to search and annotate said imagery into usable forms of data. In rural areas, where infrequent collection leads to imagery years older than in urban cores, such imagery or derived data is rendered less usable due to uncertainty around date of collection.

To address the fundamental need for governments to catalog building, road, and other land assets in a timely way at scale, Orbital Insight has developed a land use classifier algorithm leveraging Planet’s medium-resolution imagery. Planet images every region of earth multiple times per week. By ingesting this imagery into our newly-released GO platform, Orbital Insight produces land use maps that are uniformly current across any area analyzed, regardless of urban/rural breakdown. Using proprietary computer vision techniques, users may create an inventory of buildings, roads, and other land use classes at municipal, county, state, or even country-wide scale.

Case Study: Updating Buildings and Parcel Data in Northern California

We will go through a practical use-case that demonstrates the land use algorithm as a foundational data source, and specifically its ability to contribute towards land management priorities. In this case, we combined Orbital Insight building polygons from April 2019 with open-source parcel data in Northern California. Our goal was to assess the currency and validity of parcel records, specifically those marked as “Vacant” plots and “Tree Protection Zones” (TPZ). Then we imported this data into an Esri web app, providing an intuitive self-service user interface for exploring the results of that analysis.

Medium-resolution land use classifier run in Eureka, CA (Apr. 2019). The results are visualized in our GO Explore tool.

Step One — Run Orbital Insight Land Use Classifier:

Land use boundaries for Northern California, outlined in GO Create where users can geofence their own areas for processing and analysis.

The project was initiated one afternoon, and results were ready for analysis the following day! Our algorithm selected 831 individual scenes to ensure cloud-free imagery for the entire area was available. Using imagery aggregation techniques, we scanned, annotated, and combined outputs of each individual scene to create a finished, high-accuracy, high-resolution dataset for the region nearly overnight.

Users may take a first pass analysis at the land use classifier results in the GO Explore Visualization Platform Tool, or leverage GO’s download functionality and/or API. We chose the latter and will retrieve those results in a structured geospatial data format for additional analysis using third-party parcel data.

Step Two — Analyzing Building Footprints against Parcels Boundaries:

Preparing Parcels Data

Intersecting Orbital Insight Land Use Data

Tabular representation of example parcels records after we have spatially intersected them with our newly derived land use data. In addition to its original records, each one now also contains a current accounting of building and road footprint inside its boundaries. The highlighted entry was listed as “Vacant” but contained buildings, hence the alert_status is changed to “yes”.

Step Three — Making Sense of the Results

Interactive Dashboard

For the entire 10,000km2 area, 2,332 parcels contain buildings that are labeled as Vacant (orange) or Vacant & TPZ (red). A user can zoom in and pan around to explore every single one of these, along with the building outlines. The informative widgets on the left side will adjust accordingly.

This approach flags Vacant and TPZ parcels that Orbital Insight has identified buildings in. We can also render the original Orbital Insight land use data so that users may see the actual buildings detected across the region — as seen in the two images below. Visualized over flagged parcel information, these can provide a starting point for further investigation by local authorities in the event that the building is unregistered.

A commercial and industrial area on Eureka, California’s west side. Parcels are marked orange if any portion of a GO Platform building detection intersects a Vacant parcel boundary.

Repeatable and Scalable Analysis through Automation

  • Parcel Data Collectors: Inform parcel updating campaigns for local governments or private companies tasked with maintaining such datasets.
  • Tax Collection and Law Enforcement: Identify structures that may have been previously unknown to local authorities. Prioritize future assessment efforts over large areas.
  • Land Conservation Management: Gather the latest data on development in protected areas.
  • Land Management and Trend Analysis: Expand to layer in additional parcel record categories and Orbital Insight analytics (cars, trucks, airplanes, mobile phone geolocation data, and more). Conduct multivariate analyses to better understand development trends and inform policy decisions accordingly.

Email publicsector@orbitalinsight.com for additional information or to set up an introductory meeting.

Back to Blog