PostGIS

Everyone's finding out that Postgres is the database of the future, and that's the way to go.

And Postgres, of course, feeds into PostGIS.

So understanding the spatial needs.

And if you take PostGIS and you put it up against Oracle Spatial, you'll find that PostGIS is more powerful than Oracle Spatial.

But as everyone finds that out, they're thinking, well, maybe I need to migrate from your Oracles and move over to Postgres so that I have that full power backed by PostGIS as well.

So thinking about that, you'll find this trend happening across many companies today.

Postgres is used in all of your major platforms.

So I got a phone call one day, and the services that verify your physical location when you log into these betting apps, those services are obviously spatial.

This is the agency that's managing that.

And it was built using open source stack, Django, PostGIS, MapServer, Leaflet.

And the original version was built quite a while ago, even before my time at GreenInfo.

So back into the... 2012 or so.

How did we update those map descriptions? Well, there's not a lot of magic here.

It's really quite simple.

We used PostGIS on the back end to intersect key layers with the leaflet bounding box.

And then we returned that information to the front end.

Okay, so back to our enterprise environment.

Our company itself, heavy users of PostGIS, hundreds of customers equals hundreds of PostGIS databases, both cloud-based, on-premises, et cetera, et cetera.

And OpenLayers is our workhorse database.

interactive map, JavaScript library, etc.

And of course I was going to make millions of dollars, of course.

And so we had this, this is where the journey began with grandiose dreams.

And now this is Postgres, PostGIS 2007.

Now, of course, today we have a very different vibe.

So I think the education aspect was pretty good.

We started designing more tables on our own and getting more comfortable with the geospatial functions in PostGIS.

So you can see here, you probably can't see because it's really small.

So this is just our custom study area that someone drew.

So one of the big improvements we made is we started storing all of this in a table with a geometry and this ID.

So now we're mostly just sending the ID back and forth and telling our tile server what to draw.

And if we need to do any processing with it, we just say, hey, do it in the PostGIS world and then draw your results from our tile server.

So this one for example, this is the custom drawn area and you'll see in the video, it's just running ST intersects in the background and it's pretty quickly finding only those, or ST intersection, it's finding only those shapes that we want to draw.

Welcome to Cloud Optimized Shapefiles, yesterday's format of tomorrow.

So a year or so ago, my company was having an internal summit, and I was going to give a presentation on PostGIS as the resident expert.

And I did as you do, and I immediately went to Paul Ramsey's site.

I was trying to find his article on SC subdivide, because I found it pretty useful.

I have left some of these build options in the command line options so you can try some of the things that are not good ideas.

But it turned out the default one was the best.

So in conclusion, the PostGIS presentation was never completed.

But we do have a Cloud Optimized Shapefile library on MTM, which does have a command line tool for adding Quix files and reordering based on Quix files.

And I will mention here that it is ideally suited for large complex open data programs.

When you install GeoNode, you're installing PostgreSQL or PostGIS as a database, GeoServer as the data server, and MapStore as the front-end visualization analysis package.

Django is the engine that ties it all together and manages user authentication and the overall web application structure.

Jeff Terrell, Geo server map store and post just obviously can be used on their own separately.

What is Core Spatial? It's an AI-powered, open-source, geoin platform that's built specifically for federal deployment.

So it's entirely built on OzGeo technology, primarily around GeoServer, MapStore, and Postgres and PostGIS.

It has an MCP and TypeScript abstraction layer.

in front of it to enable agentic AI workflows.

There's a great blog post from Martin Davis that summarizes how this works.

This is the same functionality that's pretty recently introduced in PostGIS, too.

It's the same stuff on the back end.

There's also a simplify coverage tool.