Sid

We must stop selling files: The case for streaming tensor... Rasteret sped up geo image reads by 10x. But users still spend 80% time doing ETL before even feeding GPUs. Its like having a fast car on roads full of potholes. We investigate why, and share our solu...

Broadcast to sell, buyers see the feed and stream the pixels they want not files, monitor and monetize each byte.

If you are building for GeoAI, your data infrastructure shouldn't be your product, focus on your science and logic. Read our full blog here: blog.terrafloww.com/streaming-te...

Introducing Rasteret SDK & Terrafloww Platform.
We are killing the "search, discover, and file download" ceremony.

• Rasteret SDK: Stream EO images from S3 to GPU.
• Terrafloww Platform: Data Discovery & Metering

Think "YouTube for Geo Imagery"

Defense "ate" everything in geospatial in 2025. They bought proprietary systems that just work. The commercial EO market is left with "Download to Desktop" buttons and draw polygons on map to get data.
Fresh out of the AWS Space Accelerator, we're launching the counter-move. 🧵

⚡️Quick stat, Rasteret library gets 1 year's worth of #landsat satellite's #NDVI time series for a small farm is less than 4 seconds!

Blog - blog.terrafloww.com/rasteret-a-l...

Code - github.com/terrafloww/r...

#geospatial

Thanks to @maxlenormand.bsky.social , Soumya ranjan from @developmentseed.org , my colleague Gajesh Ladhar from Satsure , for thier early feedbacks.

GitHub - terrafloww/rasteret: A library for fast reads of Cloud Optimized Geotiff satellite images, using GeoParquet as COG metadata cache A library for fast reads of Cloud Optimized Geotiff satellite images, using GeoParquet as COG metadata cache - terrafloww/rasteret

Been working on 'Rasteret' since the last blog I wrote, its out now as an early release.

More details - blog.terrafloww.com/rasteret-a-l...

Open to feedback and contributions, there is much more exciting work to do!
github.com/terrafloww/r...

#geospatial #cloudnativegeo #opensource

Sorry to hear about this. Take care.

True totally agreed that its tough. But im wondering how PyPI gets funded by donations from both community developers and companies like Meta and so on, i guess it just due to sheer numbers? Geo is pretty small compared to general Python.

It doesn't need to another private company setting it up and owning it. It can be truly open and community + private industry funded. Unlike PyPI which hosts full libraries, the geo dataset registry doesn't need to keep STACs or any datasets inside it. A few Postgres/ES instances might be enough.

I mean PyPI registry's framework/design and governance is close to what can be done for geo as well.

I think PyPI is interesting. Its completely funded by donations. Controlled by python software foundation. Also, since I thinking there will just be pointers to STACs and Non-STACs I don't think it will be costly to maintain.

Similar to PyPI, data producers can push just metadata of datasets in toml files, to a central registry. It would only contain a "summary metadata" of entire STACs, parquet/csv files, APIs like OSM. With thier total bbox, quality metrics etc. A registry like this can be more easily governed i feel?

Christopher Beddow on LinkedIn: #gischat #maps #api #data #geospatial | 14 comments A universal map data API aggregator? Does it exist? I saw this for flight and travel APIs and some others. You can subscribe to this service, and it acts as… | 14 comments on LinkedIn

@cbed.bsky.social had put up a question on LinkedIn about finding all datasets STAC or not. And his idea was an aggregator of datasets in YAML. Similarly I feel PyPI registry is a good thing to emulate, with geopip to get data and pass it to pystac/duckdb/requests based on details sent via toml

A cropping of the immigration flow chart that reads “Born in India? Y -> DENIED”

I know it shouldn’t but this bit is so petty it made me lol

The completely open source code of iceberg and the basic REST catalog still provides great features for most people. Read/filter, write/append cloud based data (usually parquet) using just Pyiceberg. Merge/update rows are possible via Trino/Spark engines, it should come soon to Pyiceberg as well.

Feels good to have #geospatial blog on Hacker News front page!

Chime in, on the comments if you wish!

news.ycombinator.com/item?id=4241...

As I get closer to releasing the open source library which I will call, "Rasteret" , short for raster-retrieval, I reflected on the blog I published last week, and edited it to include some more details regarding compute and its code samples.

Check it out and do share your thoughts!

#geospatial

Could you add me please? Thanks

Efficient Cloud-Native Raster Data Access: An Alternative to Rasterio/GDAL A journey of optimization of cloud-based geospatial data processing. Introducing a new approach to raster data access, harnessing the power of STAC GeoParquet and cloud-native workflows to push the bo...

We discovered a 5x faster approach to querying raster time series data at Terrafloww.

Inspired by #cloudnativegeospatial (GeoParquet, COG, GeoArrow), we built a novel query method.
Read more & share your thoughts!

blog.terrafloww.com/efficient-cl...

#geospatial #cloudnativegeo