March 2023 Summaries

CARTO has achieved the Google Cloud Ready - BigQuery designation, a validation program that verifies partner integrations into Google Cloud BigQuery. This achievement demonstrates CARTO's cloud-native Location Intelligence platform meets the required functional and interoperability standards for seamless integration with BigQuery. The designation provides customers with confidence in CARTO's products working seamlessly with BigQuery, unlocking geospatial capabilities directly within the data warehouse. CARTO's expertise with BigQuery enables users to interact with location data hosted in BigQuery, eliminating ETL complexity and scalability limits for geospatial workloads.

GWR is a type of spatial statistics tool that quantifies the strength of relationships between target and correlation variables, adding more insight to analysis and helping to ascertain if trends are global or local. It performs a local least squares regression for every input cell in a continuous grid, with weights assigned based on a user-defined kernel. The output is a coefficient variable for each cell, indicating positive or negative relationships between the coefficient and target variables. GWR can be used to analyze various use cases such as traffic accident rates, CPG revenue, and house prices. In this tutorial, GWR is applied to San Francisco's street trees dataset to examine the relationship between demographic variables and tree count. The analysis involves creating a spatial index grid, aggregating demographic data, and running GWR on the compiled data. The results show positive correlations between median income and tree count in certain areas of the city.

Optimizing location planning for wind turbines is crucial to meet the global goal of reaching net zero by 2050. A site selection process must balance remote but not too remote locations, proximity to residents and infrastructure, high wind speeds, and ecological sensitivity. The guide uses boolean and ranked criteria to filter potential sites in the north-west USA, considering factors such as powerline proximity, highway accessibility, protected areas, and population density. The analysis utilizes CARTO Workflows, a visual low-code analytics tool, to rank locations based on average wind speeds and population size within 18.6 miles of each cell, resulting in a suitability index for wind farms, with areas like Central Montana emerging as promising locations.

g-Xperts, a Vienna-based company, has developed an advanced risk exposure solution for a leading European insurance company using the CARTO cloud native platform to predict risk and natural peril exposure across Europe. The application brings together 170 spatial datasets to help underwriters assess potential risks and set optimal policy premiums for key commercial assets. To address scalability issues, g-Xperts redeveloped the application with Google BigQuery and the Analytics Toolbox, unlocking advanced analysis functions and ensuring performance levels scaling with increased usage and data complexity. The migration to a cloud-native platform was pain-free, thanks to CARTO's support team, allowing g-Xperts to provide an unrivaled level of service while remaining at the forefront of the geo-applications industry.

The Spatial Data Science Bootcamp is a new event concept designed to equip data scientists and analysts with geospatial skills and knowledge to apply in their day-to-day work, following the success of previous conferences and workshops. The first bootcamp was held in Central London and featured 7 experts from across the geospatial industry and academia, covering topics such as spatial data science foundations, data architecture, Python, and data storytelling. Despite only 50% of participants having received formal training, the event highlighted a need for learning opportunities in the geospatial space. The bootcamp was part of CARTO's State of Spatial Data Science report, which found that there is a real need for learning opportunities in geospatial analysis, with many organizations lacking significant spatial analysis experience. The success of the first bootcamp has led to future events, including a New York edition on March 23rd and a London conference on May 17th and 18th.

SQL is not just for managing and manipulating databases, but also extends to spatial data analytics and visualization. It allows for the creation of 5 types of spatial data visualizations: thematic maps, dasymetric maps, dot density maps, spider maps, and great circles. These visualizations can help combat area bias in maps, provide a more representative view of data distribution, and enable modeling of individual data points. SQL's query-based nature allows for the creation of temporary subsets of datasets, eliminating the need to generate new files every time changes are made. Spatial indexes like H3 and Quadbin can be used to create regular grids, reducing bias and simplifying interpretation. Various SQL functions, such as ST_INTERSECTION, ST_ISEMPTY, ST_GENERATEPOINTS, ST_MAKELINE, and ST_GREATCIRCLE, enable the creation of these visualizations. These capabilities make SQL a powerful tool for spatial data analysis and visualization, allowing users to easily create next-level data visualizations.