Notes from presenting at FOSS4G-NA 2025

[gadomski]

@wildintellect and I presented this workshop during FOSS4G-NA 2025 in Reston, VA. Here's some notes from the experience:

• Overall, the materials were excellent, the deployed resources worked as expected, and everyone was able to step through the notebooks without error. Well done 👏🏼
• The most dramatic impact was the titiler demo (surprise, surprise). If folks hadn't seen it before, they were amazed
• I think we should shift away from using the E84 Sentinel 2 COGs if possible ... while there's nothing wrong with using them per-se, IMO it just complicates the narrative during the workshop. It'd be better (IMO) to use an authoritative source (USGS, ESA) or, failing at that, our own eoapi-devseed (though there might not be enough stuff in there to be interesting?).
• Even though I'm on the record as being a bit bearish on stactools-packages, the stactools-packages section in notebook 1 was impactful — we received a question about converting "authoritative metadata" to STAC, and that section helped answer the question.
• The vector portion was a bit tricky. Folks are often already familiar with vector services, and I found it hard to answer the question "why would I use tipg instead of my existing setup"?
• There were enough PostGIS-enabled folks that I maybe could see the benefit of adding an optional section that explores pgstac just a bit, e.g. showing the table schema, calling a function or two. This was also FOSS4G-NA so this crowed was probably more PostGIS-friendly than most will be.
• At the end of the workshop, I received the question "what would be a good next step?" Maybe a small notebook 6 could answer that question?

[hrodmn]

Thanks for the feedback from the FOSS4GNA workshop @gadomski!

I think we should shift away from using the E84 Sentinel 2 COGs if possible

Maybe the USGS Landsat COGs would be a good replacement? I went with the E84 source because it had a few nice features:

• the files are COGs!
• located in us-west-2 (proximity to the workshop eoAPI deployment, but I guess we could deploy it anywhere)
• existing STAC metadata

The vector portion was a bit tricky.

Yeah, this part of the workshop is the part I had the most trouble putting together because I have limited experience with serving vector data. I think the short answer to why tipg exists is that it is a FastAPI app that serves an OGC-compliant Feature API. I know there are other feature serving applications out there but none in the FastAPI flavor.

At the end of the workshop, I received the question "what would be a good next step?" Maybe a small notebook 6 could answer that question?

This might be where we include some deployment details! e.g. go clone eoapi-template, make some configuration updates, deploy to your AWS account. Discuss approxmate costs associated with the CDK deployment, etc.

[gadomski]

the files are COGs!

Yup, this is the big win and the reason why it's such a popular resource. I agree w/ the move to USGS Landsat, other than being requester pays they fit all those other criteria (https://registry.opendata.aws/usgs-landsat/).

I think the short answer to why tipg exists is that it is a FastAPI app that serves an OGC-compliant Feature API. I know there are other feature serving applications out there but none in the FastAPI flavor

Makes sense. This is a finer point that (I think) would have been lost on most of our workshop participants. I think it's only a small subset of people that want to customize eoAPI components (which is where the FastAPI bit would be relevant).

This might be where we include some deployment details! e.g. go clone eoapi-template, make some configuration updates, deploy to your AWS account. Discuss approxmate costs associated with the CDK deployment, etc.

Yeah, I ended up pulling up the CDK and k8s repos and just showing them to folks, and I saw them copying the urls. So I think explicit call outs to "how do" would be good.

[wildintellect]

Also one bug of our causing we should fix - when we locked the example dates, forgot to update the later queries to make sure they were within that date range. I'll do a PR with that fix in a few minutes.