geospatial

Spatiotemporal queries on content-addressed data systems like IPFS would be a big unlock for the decentralized geospatial — it'd enable time- and location-aware fetching of satellite imagery, vector data, time series data. 

EASIER Data Initiative published some research into querying spatial data on IPFS using geohashes a few months ago 🌐

Spatiotemporal queries on content-addressed data systems like IPFS would be a big unlock for the decentralized geospatial — it'd enable time- and location-aware fetching of satellite imagery, vector data, time series data. 

EASIER Data Initiative published some research into querying spatial data on IPFS using geohashes a few months ago 🌐

https://easierdata.org/notebooks/geohash-ipfs-blog-post

I haven't looked too deeply into it, but I *think* it would be quite simple, just swapping out the `https://url.com/{z}/{x}/{y}.png` for the corresponding CID under the hood.

One advantage here would be that you could point existing mapping libraries to a decentralized back end with no changes to the rest of the code. (Overall I believe that the path to adoption of decentralized geospatial tech means requiring minimal behavior change on the part of geospatial devs ...)

So what? 

One very compelling use case that Andrew Hill from @tableland / Textile thought of: locally-stored mapping data for disaster resilience.

The thinking:

- High resolution (i.e. very zoomed in) map tiles are typically only loaded and used by people in the local area relevant to the tiles
- Requiring people in remote areas to fetch data from data centers thousands of miles away creates a risk if their internet connection is severed by a natural disaster

- Having high resolution tiles cached in the same geographic area (which is likely, as others in that area also use those map tiles) provides an opportunity to retain access to the data, even if connection to the broader internet is disrupted
- This only works if tiles are content addressed, not location addressed ...

What other data is most relevant to geographic areas? Phone books / contact lists come to mind ... these are instances where geographically local data accessible via content addressing could be useful.

(Also! Note that the zxy -> IPFS idea is not the same as the geohash IPFS lookup! It's similar, though, because it would require the same construction of a spatially indexed directory system with corresponding CIDs.)

(It's been a while, but if I remember correctly Mapbox tippecanoe has similar logic — chunking up a spatial dataset into vector tiles placed in a similar z/x/y folder structure ... https://github.com/mapbox/tippecanoe)

Something I've been wanting to build for a while: an adapter to a library like Leaflet or Mapbox GL JS that would wire up zxy tiling systems (both raster and vector tiles) to tiles stored on IPFS.