posted 2 hours ago
I have been slowly coming to the conclusion that society will happily spend billions making it easy to buy things, pave things, drain things, insure things, speculate on things, and destroy things. But if you want to understand the five acres you actually live on? Good luck, nerd.
There is an absurd amount of public information about land: soil surveys, elevation data, watershed maps, vegetation records, climate data, native plant databases, and decades of government research. The problem is that it is scattered across enough different systems to make a normal person regret developing curiosity.
So I started building Groundskeeper. The basic idea is simple: draw the boundary of your land, and the computer should do the boring archaeological dig through public environmental data for you.
Not because computers understand land better than people do. They absolutely do not. A computer has never stepped into a hidden wet patch, watched runoff cut through a tractor track, noticed a plant thriving where the map says it shouldn't, or discovered that the previous owner buried something inconvenient exactly where your elegant design wanted a pond. But computers are very good at reading 40 boring datasets without becoming spiritually exhausted.
Groundskeeper is gradually learning to assemble terrain, slope, soils, watersheds, probable water movement, vegetation, native plants, erosion concerns, access, infrastructure, and possible restoration or water-harvesting opportunities. I do not want it producing magical "dig a swale here" prescriptions from a raster and a prayer. I want it saying things like: this area has a meaningful catchment, these soils behave differently, this patch deserves inspection, these plants are plausible here, and something interesting is happening at this location so go look at it.
Basically, I want the computer to make walking the land more productive, not replace walking it.
Long term, I'd like Groundskeeper to become a cheap ecological planning and field tool for small landowners, restoration projects, native plant people, and citizen scientists. Right now, though, I need real properties to break it on.
If you have a property you're willing to use as a test case, send me the location or property boundary and a sentence or two about what you're trying to understand. I'm especially interested in sites with drainage, erosion, ponds, slopes, native planting, restoration, access, or "I just bought this land and have no idea where to start" problems.
I'll run a handful of properties through Groundskeeper and share what it finds, along with where the data seems useful, misleading, or incomplete. The goal is not to prove the software is brilliant. The goal is to find out where it lies before anyone trusts it.
Also curious: what information do you wish you'd had when you first started working with your land, and what would you never trust a computer to tell you without boots on the ground?
