11-29, 13:30–13:50 (Asia/Seoul), Taepyeong Hall
Working with data has many challenges on it’s own and when you add a geospatial factor into makes it even harder. In my belief data never lives in isolation, it always exists in relation to something (sorry NoSQL) and PostgreSQL provides a great starting point as our not just geospatial database but database for everything.
In this talk we’ll talk briefly about long history of postgres and how it’s ecosystem can help you build a database for all your needs. In particular we’ll look at extensions like:
- PostGIS (obviously)
- pg_raster (for handling rasters)
- timescaleDB (for handling time series data)
- uber h3 (h3 indexing for your geospatial data)
- ogr_fdw (OGR FDW allows you to connect to any OGR supported data source.)
- pg_eventserv, pg_cron, pg_ivm and many more will be referenced
Moreover we’ll look at how you can combine some extensions like (postgis[spatial] + timescaleDB[temporal]) to create something that fits your use case.
Now we have designed our database but it’s time to deploy it on the cloud, let’s look at available solutions which includes from self-managed to service based and everything in between.
- Managed services like - AWS RDS. Aurora, CrunchyData, EDB, Timescale
- Deploying your own using Linux Instances
- Deploying on K8s clusters using postgres operators
We’ll use OSM data as our sample datasets plus more as needed.
I have been honing my skills in the geospatial domain, gaining diverse experience in Climate tech startups, Agritech solutions, and more.
My experience extends to working with satellite data (including Sentinel and Landsat), geospatial data modeling, and handling large datasets at scale in the cloud using Docker, Python, S3, etc
I am most interested in building a generic spatial-temporal database that can handle a wide variety of data and use cases.