Fast Spatial Data Processing & Visualization at Scale: Insights from a Telco Use Case

Turn your spatial SQL queries into dynamic, interactive maps. Discover how data teams can achieve unmatched price performance, streamline visualization, and leverage a modern lakehouse architecture for actionable geospatial insights using Wherobots and Dekart.

Using a telco use case as an example, you’ll learn how to:

Run complex spatial queries directly on cloud-native Iceberg tables using Wherobots.
Instantly visualize cell coverage, signal drop zones, or congestion patterns in Dekart.
Layer in cluster analysis (e.g., Getis-Ord Gi* statistics) to detect statistically significant hot and cold spots.
Iterate quickly, comparing performance before and after tower optimizations.
Easily share insights with planners, engineers, and executives, no GIS software required.

This session is ideal for those who work with large, dynamic spatial datasets and turn those SQL queries to shareable interactive maps for further data exploration. No exporting data. No manual mapping. Just fast, repeatable insights from data about the physical world.

0:00 – Introduction
2:02 – Challenges in geospatial data processing that Wherobots aims to solve
7:12 – History of Apache Sedona and how Wherobots' open core platform fits into it
10:48 – Spatial compute engine performance comparison: WherobotsDB vs. Apache Sedona vs. BigQuery vs. Snowflake
14:21 – Spatial Data Lakehouse Engine: How it works
17:27 – Raster inference for geospatial AI
26:07 – Scalable geospatial analytics for telco and internet providers
28:44 – Getting started with Wherobots interface
30:21 – [Demo] Quality of Service assessment: Hot/Cold spot detection – processing 20 million GPS points in under 1 minute at a cost of ~$3
32:11 – Generate H3 grid cells and aggregate for analysis
34:30 – Getis-Ord Gi* analysis: Hotspot detection for download and upload latency
39:08 – Dekart integration with running cluster
42:18 – Overview of KNN joins
48:36 – Q&A and Wrap-up

Try Wherobots and get started here: https://bit.ly/3FQ5N95