March 2022 Summaries

Spatial SQL is a query language that enables users to work with geometry and geography data types, allowing for faster geospatial data processing and analysis, as well as support for spatial modeling and machine learning. Unlike traditional SQL, Spatial SQL has deeper capabilities that assist in areas like analyzing points, lines, and polygons, and can be used to analyze data types such as addresses, place names, latitude and longitude coordinates, and more. The language provides numerous advantages, including increased efficiency in analytics workflows, support for cross-functional collaboration, and accessibility to the wider analytics community. Spatial SQL is typically used by GIS analysts, data scientists, developers, and teams working with spatial data in SQL-based databases or warehouses, but its adoption is growing among roles associated with less technical departments like marketing, business intelligence, and operations. The language has a relatively high learning curve, but the return on investment makes it worthwhile for those willing to learn it.

The introduction of GeoParquet aims to standardize the storage of geospatial vector data in Parquet, a popular columnar storage format, with the goal of enabling interoperability between different cloud data warehouses and computing engines when using geospatial data. This initiative seeks to unlock spatial analytics for the growing enterprise software segment, which is driving a huge influx of new users to the analytics field. GeoParquet builds upon Parquet's advantages, such as its column-based format, which works well with denormalized datasets and enables efficient querying. By defining a common way to encode and describe spatial data, GeoParquet aims to address interoperability problems, making it easier to create and share geospatial data, and ultimately realizing the vision of a cloud-native data lake architecture without data copy. The initiative has received support from major clouds and products in the ecosystem, and is being discussed at the Open Geospatial Consortium, with CARTO joining as a sponsor and contributor to facilitate interoperability within the spatial analytics ecosystem.

The article investigates spatial patterns of coffee shops across New York City, analyzing caffeine deprivation through Location Intelligence. It explores how data-driven site selection can help coffee brands and retailers unlock their potential by identifying areas with a large population but limited coffee shop provision, such as Arden Heights and Grasmere on Staten Island, and southeast Brooklyn and Queens. The analysis uses the SafeGraph Core Places dataset to identify coffee shops within New York City and the H3 hexagonal grid for spatial aggregation. The study creates a heatmap showing areas of caffeine deprivation, which can be used to inform site selection and improve customer experience. The findings highlight the importance of location data in avoiding million-dollar mistakes in site selection and provide insights into the utility of provision analysis data beyond site selection.

The International Women's Day report maps financial inequality across the globe by gender, using visualizations developed with the CARTO Builder tool and its cloud-native Location Intelligence platform. The analysis highlights significant differences in access to financial services between men and women worldwide, with most countries showing greater male participation. Inequality is particularly prevalent in regions such as Egypt, Jordan, Saudi Arabia, and certain parts of Asia and Africa, where women have lower rates of receiving a wage, accessing a bank account, owning a debit or credit card, borrowing money for medical purposes, making savings, saving specifically for old age, having a housing loan, accessing emergency funds, and sending/receiving digital payments. Despite these disparities, some countries exhibit greater equality in certain areas, such as the "borrowed for medical purposes" indicator, where more women participate. The report emphasizes the importance of data visualization and spatial analysis in understanding and addressing financial inequality, and offers resources and support through CARTO's Spatial Data Catalog and grant scheme.

The Analytics Toolbox for BigQuery has been enhanced with a new set of retail-specific functions, including revenue prediction and white space analysis. White space analysis solves the question of where to open a new store by identifying locations with high expected revenue and meeting other business criteria. To complete end-to-end analysis, three steps are necessary: data preparation, model training, and finding whitespace areas. Data preparation involves enriching the area of interest with retailer and competitor data, as well as data from the Data Observatory subscription. Model training builds a revenue prediction model using Boosted Tree Regressor, while finding whitespace areas uses the FIND_WHITESPACE_AREAS procedure to identify locations within the area of interest where a new store would perform best. The procedures are now available to CARTO users and will soon power the white space analysis engine integrated in CARTO's Site Selection application.