Fused

???Running on the fly Machine Learning Inference with Fused Jeff Faudi ???????? shows us how he uses Fused to streamline computer vision inference on satellite imagery using Fused ???Read: https://lnkd.in/gv7tt3Bm?Watch: https://lnkd.in/g85tauPh

Query Foursquare's 100M OS places in a one-liner with DuckDB without downloading! If you'd like to see for yourself what kind of entries you can find in Foursquare's dataset, check out this Hugging Face dataset. Link: https://lnkd.in/dk84BKzz Based on Fused's preprocessed geoparquet files (thanks a lot!) I created a single 10Gb file that you can query with DuckDB from your system without downloading it thanks to the magic of httpfs (http range requests). It's as simple as installing DuckDB and running this one-liner: ????????????.??????(??"?????????????*??????????'????://????????????????/????-????/????????????????????_????????????_????????/????????????????????_????????????.??????????????'????????????????????????????????'%????????????%'?") On my system it takes 70 seconds for a query. If you download the file instead and query it locally, it's much faster. On my system (M3 Max), it takes only 5 seconds to query the file! ????????????.??????("?????????????*??????????'????????????????????_????????????.??????????????'????????????????????????????????'%????????????%'?") fyi: As many people have pointed out, there's indeed a whole lot of geospatial defamation in there which - in my opinion - should absolutely have been filtered out.

Vegan Restaurants and Yoga Studios in San Francisco - FSQ Places I made a 500 meter buffer around vegan restaurants (purple) and overlayed yoga studios (yellow points). This is sort of a joke but you can really see where where the market is saturated and where there's oppurtunity for new establishments. Fused

Drinking and Dining - Foursquare Places and Overture Builidngs This was an H3 join at resolution 11, there is still a lot that shows up at 12. You can see in Brooklyn the establishments can generally be found on avenues. This is such a rich and very precise dataset. I'm looking forward to dining in. Thinking about visualizing competition zones. Fused

Flood Analysis with Sentinel-2 and Overture Maps, DeLand Florida I've been experimenting with the Fused satellite UDFs and I found that if I filter the Near Infared (NIR) band from Sentinel-2 enough I can isolate water. I turned the band into blue hexagons that are seen here and joined them with Overture buildings (orange) to see if the water intersected them. I wanted to test my threshold so I looked at the Spring of 2024 in the same area that was fairly dry against October when Florida was hit by a series of hurricanes. You can use methods like this to see if structures are affected in environmental disasters. For those of you in insurance, each of these buildings has an address from Overture.

#30daymapchallenge day 22: 2 colors Mapping places in the Foursquare open source POI dataset with "San Francisco" (red) or "Oakland" (blue) in the name. There's now a red (SF) POI in the middle of Oakland airport. ?? Blog post about accessing the data: https://lnkd.in/gkFamrd9 All data hosted by Source Cooperative, made in Fused. The land data comes from Overture Maps Foundation.

Heatmap of "Business and Professional Services" in NYC from Four Square's Places dataset on Fused.

Access the New Foursquare POI dataset in a single click, all from your browser ??

???How Pachama creates maps on-the-fly with Fused Andrew Campbell, Senior Software Engineer at Pachama shows us how he’s been using Fused as part of their Land Suitability Tool ???Read: https://lnkd.in/gD_KvZ8p ??Watch: https://lnkd.in/gGDUGMQK

In this article I break down the Pittsburgh post from earlier, looking at the H3 DuckDB functions h3_grid_disk h3_polygon_wkt_to_cells and argue why these can be used for dynamic and adaptive data visualization.