A field manual for engineering production spatiotemporal & movement-data pipelines in Python.

spatiotemporalmovement.com →

Spatiotemporal Movement is a deep, continually growing engineering reference for the people who build mobility analytics: mobility data scientists, urban analysts, Python GIS developers, and logistics/transportation engineering teams. Every page is a self-contained, executable guide — grounded in coordinate and temporal precision, production-ready Python, and the messy reality of debugging real-world movement data, not theoretical GIS overviews.

Each guide ships with complete, runnable code (typed signatures, explicit edge cases, vectorized patterns that scale past notebooks), hand-authored diagrams, validation strategies, and a tight web of cross-links so you can move from a symptom to a fix in a couple of clicks.

The material is organised into four sections, each broken into narrowly scoped topics:

• Spatiotemporal Data Foundations & Structures — trajectory object design, coordinate reference systems, GPS precision & error handling, sampling-rate optimization, time-series synchronization, and spatial storage formats (GeoParquet).
• Movement Pattern Extraction & Trajectory Analysis — trajectory segmentation, stay-point detection, speed/acceleration profiling, directionality & turn analysis, and movement similarity & clustering (Fréchet, DTW).
• Temporal Aggregation & Window Mapping — dynamic time-binning, rolling statistics, gap filling in sparse trajectories, seasonal/cyclical alignment, and cross-border time-series synchronization.
• Spatial Indexing & Density Aggregation — discrete global grids (H3 / S2 / quadkey), map-matching to road networks (HMM), origin-destination flow matrices, and kernel density surfaces.

• Handling GPS Drift in Raw Trajectory Logs
• H3 vs S2 vs Quadkey for Mobility Density Grids
• Feeding Cleaned GPS Output into an HMM Map-Matcher
• DBSCAN for Stay-Point Clustering in Python
• Kalman Filter Gap Filling for GPS Dropouts

pandas · polars · geopandas · shapely · pyproj · movingpandas · scikit-mobility · scipy · scikit-learn · h3 · s2 · osmnx · rasterio · pyarrow / GeoParquet · duckdb

Every code sample follows the same discipline the guides preach: all distance, speed, and acceleration math is done in a metric projected CRS (never raw EPSG:4326), timestamps are UTC-first, and window boundaries are deterministic.

A static site generated with Eleventy, authored in Markdown with inline, hand-authored SVG diagrams and structured data (JSON-LD). It is deployed to Cloudflare Pages.

npm install
npm run build     # build the static site into _site/
npm run serve     # local dev server on http://localhost:8080

Every page is held to a strict, automated quality bar — accessibility (WCAG 2 A/AA via axe-core), valid structured data, internal-link integrity, responsive layout, unique SEO metadata, and a Lighthouse performance budget — before it ships.

Spotted an error, or want a topic covered? Open an issue on this repository. Corrections to code, thresholds, and edge-case handling are especially welcome.

Built and maintained by Spatiotemporal Movement.