Seamlessly integrate Google Earth Engine with CARTO Workflows

At CARTO, our mission is to break down data silos and make geospatial analysis more accessible, scalable, and impactful. With our cloud-native platform, we empower organizations to turn spatial data into actionable insights - whether for urban planning, environmental monitoring, insurance risk assessment, or beyond.

One of the most powerful resources available for spatial analytics is Google Earth Engine (GEE), a vast repository of global-scale raster data, including climate and environmental variables, satellite imagery, and geophysical datasets. However, efficiently accessing and analyzing this data - especially within SQL-driven analytics environments - has historically been a challenge for users without more advanced JavaScript skills.

That’s why we’re excited to introduce a new Google Earth Engine Extension Package in CARTO Workflows - our low-code tool for automating your spatial analysis. This extension is designed to integrate GEE data seamlessly into your spatial analysis with a set of components that make it easier than ever to extract, analyze, and visualize raster data at scale. This is all without you needing to write complex queries or develop custom integrations.

With this extension, you can now bring Earth Engine’s powerful geospatial capabilities directly into your cloud data warehouse, ensuring a scalable, secure, and efficient approach to spatial analysis.

Above: NDVI analysis using the Google Earth Engine Workflows Extension Package

GEE provides an unparalleled repository of raster datasets, from historical climate data to real-time environmental monitoring. However, integrating it into a cloud-based analytics workflow often requires significant technical expertise.

With CARTO Workflows’ new extension package, you can now:

• Extract raster data (e.g., precipitation, elevation, NDVI) directly from Earth Engine.
• Summarize geospatial patterns across large datasets with scalable, cloud-native processing.
• Perform time-series analysis to track environmental changes over time.
• Seamlessly integrate insights into your SQL-based workflows without deep knowledge of remote sensing.

This means you can take advantage of Earth Engine’s vast dataset library - without the complexity of writing custom scripts or managing API integrations.

The impact of this integration extends across multiple industries, helping organizations make data-driven decisions with ease. Here are some examples of key applications:

• Insurance: Assess wildfire risk scores for insured properties using historical and real-time environmental data.
• Utilities: Evaluate flood risk for power plants and infrastructure based on precipitation and terrain analysis.
• Telecom: Extract environmental risk factors for mobile towers and antennas, improving network resilience.

By leveraging CARTO Workflows with Earth Engine, businesses can optimize risk assessment, enhance operational efficiency, and improve long-term planning. Let’s explore how!

This extension includes five components for you to use:

• Get Elevation: Retrieve elevation data for specified polygons, enabling topographic analysis.
• Get NDVI: extracts Normalized Difference Vegetation Index (NDVI) values using NIR (B8) and RED (B4) bands to monitor vegetation health.
• Get Precipitation: access historical and real-time precipitation data from GEE to assess climate patterns.
• Summary Region: send a single polygon to Earth Engine and retrieve summarized data (e.g., average wind speed over a given area). Best for small queries (<10 records), with results in under a second.
• Summary Table: process large datasets (10,000+ geometries) by retrieving summarized values in batch mode. Designed for scalability, handling millions of records efficiently.

With these components, users can extract the data they need from Google Earth Engine’s catalog - whether for environmental monitoring, disaster risk assessment, or infrastructure planning.

Let’s see this in action!

Imagine we are data analysts at a US property insurance company who insure 10,000 properties across Texas. Following a period of sustained rainfall, we have received 888 compensation claims. We’ve been tasked with identifying which of these could potentially be fraudulent and may require further evidence before being processed.

Fraud detection is a major challenge for industries like insurance and financial services, where bad actors attempt to manipulate location-based claims. By integrating Google Earth Engine (GEE) with CARTO Workflows, businesses can leverage satellite imagery and environmental data to uncover fraudulent activity with greater accuracy and efficiency.

We’ve designed this extension to be flexible, scalable, and intuitive - so anyone can use it with the following steps:

1. To access the extension, you’ll need a CARTO account - you can sign up for a free 14-day trial here.
2. Once logged in to the CARTO Workspace, head to the Workflows tab and create a new workflow with your Google BigQuery connection. Alternatively, you can use the CARTO Data Warehouse.
3. Switch to the Components tab on the left of the screen, and click Manage Extension Packages at the bottom-left.
4. Select the Google Earth Engine Extension Package (learn about the other available packages here) and select install.
5. Once the package has finished installing, you will be able to find the available components in the Components window - and you’re ready to start working with Google Earth Engine - the low code way!

To assess the likelihood of fraudulent claims occurring, we will compare the (fictional!) claims against precipitation volume and pre-existing flood risk analysis, for which we’re using the National Risk Index for Natural Hazards from FEMA. We’ll achieve this by building the following workflow.

• Step 1: Extract precipitation data from GEE: In this step, the Summary Table component from our new Google Earth Engine Extension is used to extract the maximum daily precipitation volume from the GRIDMET: University of Idaho Gridded Surface Meteorological Dataset image collection. We do this with the following custom image expression, setting a date filter as well as a bounding box to make data retrieval faster:

   
image = ee.ImageCollection('IDAHO_EPSCOR/GRIDMET')    
.filter(ee.Filter.date('2024-04-26T00:00:00', '2024-05-04T00:00:00'))    
.filterBounds(ee.Geometry.Rectangle([-106.0, 25.8, -93.5, 36.5]))    
.mosaic();
   

• Step 2: Join with flood risk data: Secondly, the Spatial Join component is used to join each claim to the FEMA risk areas, which contain both river and coastal flood risk scores.
• Step 3: Assess the likelihood of the claim being fraudulent: Using the combined inputs of precipitation volume and flood risk, each asset is assessed on how likely they are to be fraudulent. An asset that has experienced high rainfall and is in a high flood risk area would be deemed a negligible risk, whilst one with low rainfall in a low risk area would be deemed a higher risk.

Let’s check out the results!

You can explore the results of this analysis in the CARTO Builder map below! Light yellow assets are ones where the fraud risk is lower, moving up to darker red assets where the risk is higher. The asset pop-ups and dynamic widgets on the right of the map can be used to explore the results - and the factors which have been driving them - in further detail.

Insurers could reduce internal costs by greenlighting compensation in areas with low or negligible risk without requiring time-consuming manual assessments, freeing up staff time to further investigate higher risk locations.

We can make obtaining insights from our map even easier with the use of CARTO AI Agents - our AI-powered solution which empowers anyone - no matter their skill level - to derive actionable insights from spatial data using natural language prompts. Check it out in action below!

Our new Google Earth Engine Workflows extension package is enabling faster, more scalable, and more intuitive spatial analytics. We are actively collaborating with Google to further enhance these capabilities - ensuring CARTO users get the best possible access to Earth Engine’s powerful data.

🚀 Ready to get started? Watch our recent webinar with Google: Analyzing the Earth at Scale to see this in action, and sign up for a free 14-day trial to try it yourself!

With CARTO + Google Earth Engine, the possibilities for scalable, cloud-native spatial analysis are endless. Start exploring today! 🌎

Want to get started?

Start free trial