Simplifying Geospatial Analytics with the New Sedona STAC Reader

Integrating extensive satellite imagery and geospatial datasets into analytics platforms has traditionally been a complex and time-consuming effort. As the primary developer and contributor to the SpatioTemporal Asset Catalog (STAC) Reader feature, I’m excited to announce its introduction in the upcoming release of Apache Sedona 1.7.1. This feature simplifies geospatial analytics by enabling seamless ingestion and analysis of multi-dimensional datasets.

For example, urban planners assessing a city’s resilience to climate-induced flooding can utilize the STAC Reader to integrate and analyze?Sentinel-2 satellite imagery efficiently, identifying vulnerable areas and informing effective mitigation strategies. Here’s how this functionality can be implemented:

// Load STAC data into Sedona dataframe
val df = spark.read
  .format("stac")
  .option("itemsLimitMax", "100")
  .option("itemsLoadProcessReportThreshold", "2000000")
  .option("itemsLimitPerRequest", "100")
  .load("https://earth-search.aws.element84.com/v1/collections/sentinel-2-pre-c1-l2a")

// Show schema and data
df.printSchema()        

Despite the wealth of information contained within STAC datasets, professionals often encounter several obstacles when integrating them into geospatial analytics workflows:

• Scalability: Geospatial datasets, especially high-resolution imagery like Sentinel-2, are large and computationally demanding. Processing such data can lead to performance bottlenecks, extended processing times, and increased costs without adequate infrastructure and optimized workflows.
• Analytical Workflow Integration: Incorporating geospatial data into existing analytical workflows can be challenging due to its specialized nature and potential lack of institutional knowledge. This skills gap can impede the adoption of geospatial technologies and the integration of STAC datasets into existing analytical frameworks.

Recognizing these challenges, the Apache Sedona community will introduce the STAC Reader in version 1.7.1 to facilitate smoother integration of multi-dimensional datasets into geospatial analytics and queries. By supporting the STAC specification, Sedona enables users to seamlessly access and analyze diverse geospatial assets, such as Sentinel-2 satellite imagery, without extensive data preprocessing. This integration aligns with Sedona’s mission to provide efficient and scalable solutions for spatial data processing, empowering users to derive valuable insights from complex datasets with greater ease.

-- In this example, the data source pushes down the temporal filter to the 
-- underlying data source, efficiently retrieving records within the 
-- specified date range.
SELECT id, datetime as dt, geometry, bbox
FROM STAC_TABLE
WHERE datetime BETWEEN '2020-01-01' AND '2020-12-13'


-- In this example, the spatial filter is applied to retrieve records whose 
-- geometries are contained within the specified polygon, optimizing data 
-- retrieval by leveraging spatial indexing.
SELECT id, geometry
FROM STAC_TABLE
WHERE st_contains(ST_GeomFromText('POLYGON((17 10, 18 10, 18 11, 17 11, 17 10))'), geometry)
        

The launch of the STAC Reader in Apache Sedona 1.7.1 marks a significant advancement for professionals working with geospatial data. This new feature tackles common challenges related to the integration of STAC datasets, simplifying data ingestion and improving analytical workflows. As a result, users can fully utilize the potential of geospatial data, leading to more informed decision-making and encouraging innovation across various industries.