Dewey Dunnington

Dewey Dunnington

@paleolimbot.bsky.social

Scaling spatial computing at Wherobots, Inc. ApacheArrow PMC, #gischat, #rstats, ex Voltron Data.

271 followers100 following109 posts

Dewey Dunnington reposted

Replying to @kylebarron.dev

triangles "flying" from input coordinate space to output coordinate space

Dewey Dunnington reposted

📣 Just released: Version 25 of the ADBC libraries. Release highlights: C# core libraries at 1.0, Windows on ARM packages, Python 3.15 wheels, separate Go modules for core and drivers, less unsafe Rust, and PostgreSQL and SQLite driver improvements. arrow.apache.org/blog/2026/10...

Excited to announce Apache SedonaDB 0.5.0! In this release is an improved interface for reading various file formats, DataFrame unnest()ing, improved raster functionality, an improved GeoPandas compatibility layer, ever more spatial functions, and the seeds of foreign catalog integration.

View full thread

Excited to announce Apache SedonaDB 0.5.0! In this release is an improved interface for reading various file formats, DataFrame unnest()ing, improved raster functionality, an improved GeoPandas compatibility layer, ever more spatial functions, and the seeds of foreign catalog integration.

View full thread

We have an improved GeoPandas compatibility layer as well in this release...while most of the improvement you can get out of SedonaDB is by rethinking your workflow, GeoPandas is a fantastic and well-worn API. We also fixed a lot of bugs from being able to run their tests!

Excited to announce Apache SedonaDB 0.5.0! In this release is an improved interface for reading various file formats, DataFrame unnest()ing, improved raster functionality, an improved GeoPandas compatibility layer, ever more spatial functions, and the seeds of foreign catalog integration.

View full thread

I've always loved the R ecosystem's unnest() verbs, and in SedonaDB 0.5.0 you can do this in Python! In combination with the venerable ST_💩(), this can help with data wrangling of multi- geometries. (Unfortunately I forgot to add the #rstats version...next release!)

Our raster/ndarray support is still ongoing, but as of SedonaDB 0.5.0 you can read files and load pixels with support for a few real-world operations like sampling and zonal statistics. In Python you can also pass numpy arrays zero-copy into query!

Excited to announce Apache SedonaDB 0.5.0! In this release is an improved interface for reading various file formats, DataFrame unnest()ing, improved raster functionality, an improved GeoPandas compatibility layer, ever more spatial functions, and the seeds of foreign catalog integration.

Getting data into a query is important and SedonaDB has always tried to make it easy; however, this led to a lot of different functions and mental mayhem when switching between #rstats, Python, and SQL. Our 0.5.0 release routed everything through the same pathway!

Dewey Dunnington reposted

Cloud Native Geo folks!!! Check out the great talks happening at FOSS4G NA! #CNG #GISchat #EO

FOSS4G North America@foss4gna.bsky.social

🌐 Whether you're new to Cloud-Native Geo #CNG or a practitioner, there is something for you. Explore hands-on workshops & talks covering #STAC, #GeoParquet, #DuckDB, vector tiles, #Zarr, cloud-optimized point clouds, reproducibility, and large-scale data publishing.

Excited to be talking SedonaDB at Cloud-Native Geospatial Forum (CNG) forum next week, and even more excited to connect with the spatial community in person for the first time in a while. If you're attending let me know and make sure we connect!

Dewey Dunnington reposted

I cannot emphasize enough how excited I am about {geoarrowwidget} and {deckglgeoarrow}. 3.2 **million** points in 4 seconds. #rspatial #rstats

Excited to be teaching again! Sign up for the Monday SedonaDB workshop to mostly learn about how query engines/databases can make large(r) data in Python less awkward!

FOSS4G North America@foss4gna.bsky.social

Arrive in Sacramento early and dive into hands-on workshops. Here is a sample: 🔹 Reproducible geo environments with Nix 🔹 Large spatial data in Python + SedonaDB 🔹 Disaggregating commodity flows 🔹 Cloud-native vector formats #GIS #Geospatial #EO #CNG

Dewey Dunnington reposted

It looks like {geoarrowrs} is even competitive with {sedonadb}?! Do we got a new fast spatial contender! #rstats josiahparry.github.io/geoarrowrs/a...

Josiah@josiah.rs

{geoarrowrs} is very competitive with {duckspatial} and even when calculating the **correct** distances on geodesic data #rspatial

Dewey Dunnington reposted

Today, I talked about 🐘🪿⛰️, or more precisely, about PostGIS, DuckDB and Apache SedonaDB at FOSS4G Hiroshima 2026. Here's my slides. Guess which one I chose to recommend after this roulette?? 🎰 SedonaDB: Why Yet Another Geospatial Database Engine? github.com/yutannihilat...

We're excited to announce a new release of SedonaDB, and this one is a big one! Between native DataFrame APIs in Python and R, a brand new Geography implementation, GPU accelerated joins, and a raster/ND-array foundation, it's hard to know where to start.

View full thread

Finally, we've laid a foundation for raster, Zarr, and ND-array functionality primarily by GDAL! SedonaDB 0.4.0 includes basic metadata and load operations, and we're exited to build on this with the community for our next release.

We're excited to announce a new release of SedonaDB, and this one is a big one! Between native DataFrame APIs in Python and R, a brand new Geography implementation, GPU accelerated joins, and a raster/ND-array foundation, it's hard to know where to start.

View full thread

While we had some intial support from our first release, in 0.4.0 we rewrote the geography implementation to ensure full support, including Parquet pruning, and spatial joins. Buffers, dwithin joins, and interpolation of points along the shortest geographical path should now be easy and performant!

This release features an accelerated join by Liang Geng, productionizing his research using RT cores on GPUs to accelerate spatial joins. Compute-intensive joins can be many times faster when SedonaDB is built with this feature enabled!

We're excited to announce a new release of SedonaDB, and this one is a big one! Between native DataFrame APIs in Python and R, a brand new Geography implementation, GPU accelerated joins, and a raster/ND-array foundation, it's hard to know where to start.

View full thread

After writing the last releast post, I mentioned to my colleagues that it would be the last one where I show off the amazing capabilities of our engine with clunky SQL. I'm proud to share that we made this happen! You can now .select, .filter, .group_by, .agg, and .sort your Python DataFrames!

...and, of course, we didn't leave out #rstats. The 0.4.0 release features the sdplyr R package, which provides a dplyr interface covering the basic subset of transformations (including spatial joins!). Like Python, our ST_ and RS_ functions have inline docs via the generated sedonafns package.

We're excited to announce a new release of SedonaDB, and this one is a big one! Between native DataFrame APIs in Python and R, a brand new Geography implementation, GPU accelerated joins, and a raster/ND-array foundation, it's hard to know where to start.

...or maybe not that hard, since you'll have to install it to get going. We're excited to have a contributed conda-forge package to make it easier for conda users to integrate SedonaDB into their workflows.

Dewey Dunnington reposted

Do you love maps? Of course you do! With the latest release of #ggsql we now support spatial plotting directly from your database. As a cherry on top, the release also provides powerful in-layer aggregation. Read all about it on the blog

opensource.posit.coggsql 0.4.1: Spatial plotting and in-layer aggregationggsql 0.4.1 adds support for spatial plotting and data aggregation within layers. Read all about the new features in this release post.

Dewey Dunnington reposted

dbplyr 2.6.0 is here! This release leaned on Claude Code to clear a TON of smaller issues, freeing up time for the big stuff: brand-new ADBC and JDBC backends, new translations, and a new sql_dialect() to cleanly decouple connection from SQL dialect. Read more: opensource.posit.co/blog/2026-06...

opensource.posit.codbplyr 2.6.0dbplyr 2.6.0 is now on CRAN, with new ADBC and JDBC backends, IBM DB2 support, more translations, and a bunch of minor fixes and improvements.

Dewey Dunnington reposted

I'm alive *AND* I wrote a blog post?! Unheard of! Much 💖 to @abbycabs.dev, and @grafana.bsky.social's David Allen and Amanda Wagner whose FOSDEM talk's were a huge source of inspo!

stdlib@stdlib.io

For years, open source communities relied on visible friction: people got stuck, asked questions in public, and became part of a community. AI is changing that. What happens when newcomers are no longer visible—and neither side knows the other is there? blog.stdlib.io/ai-and-the-i...

Dewey Dunnington reposted

We are thrilled to announce that Positron Server is now available for academic use via JupyterHub! This gives students a robust #RStats & #Python data science IDE without needing a local install or new infrastructure. Learn more: positron.posit.co/blog/posts/2...

Can your current spatial SQL engine ST_💩()? We didn't think so! As part of our ongoing attempt to innovate on spatial SQL in Apache Sedona, this we're proposing emoji shorthands for most functions which (1) improve expressiveness and (2) helps more SQL fit on one line.

Dewey Dunnington reposted

We have summer internships y'all! Come work at Posit on the PyData, tidymodels, shiny, or Connect teams: grnh.se/tigz810a3us. You will have an awesome time, learn a ton, and help advance our open source and pro tools 🧰 #rstats #pydata

Dewey Dunnington reposted

I tend to be a CRAN-appreciator. But wow, R-Universe is such a useful project. As a developer, I can: - Experiment with new and creative approaches that would never get accepted on CRAN; - Have cross-platform binaries for my packages built within an hour (usually) of a GitHub push

Dewey Dunnington reposted

The fastest operation is the one you don’t have to do. When a database natively supports @arrow.apache.org, ADBC can speed up fetching and ingestion by eliminating costly row/column conversions. How much faster is it in practice? We ran some benchmarks to find out. Link below 👇