Every property screen is an argument assembled from records. If you cannot see which records carried a given property, you cannot judge whether the argument holds. So it is worth walking the source list category by category and being specific about what each one settles and where each one stops.
Most of what follows is public. USDA NAIP imagery, USGS 3DEP elevation, FEMA flood mapping, NOAA storm records, county assessor files and municipal permit portals are open programmes that any competent geospatial team can work from. Naming them is a statement about where the research looks, not a claim of privileged access.
Aerial and satellite imagery
Imagery is the only source that shows you the property as it physically is. Roof shape and pitch, visible obstructions, an existing array, parking and laydown area, tree canopy over a residential roof, site access and adjacency all come from here.
Three properties of an image decide how much it is worth. Resolution decides what you can identify at all. Vintage decides whether what you are seeing still exists. Season decides what is hidden.
- NAIP is flown on roughly a two-to-three-year cycle, state by state, so a given county's most recent capture may be close to three years old.
- NAIP is leaf-on and captured near midday, which means winter shading and seasonal obstruction are invisible in it and shadows are at their shortest.
- Commercial captures are usually higher resolution and more recent, which is why they earn their cost on a question imagery alone has to answer.
- An oblique view shows a facade, a parapet, or a wall-mounted service entrance that a straight-down orthophoto flattens away.
- Anything built, replaced, or removed since the flight date is simply absent from the record.
The failure mode is treating an image as current. A rank that rests on a three-year-old capture should say so in the report, with the capture date printed next to the observation.
County parcel and assessor records
Assessor data is the backbone of a property screen because it is the only source that reliably ties a geometry to an identity. Parcel boundary, land-use or property-class code, building area, year built, owner of record, and in many counties roof cover, construction type, and last sale date all come from the assessment roll.
This is also the source that covers both halves of a territory. The same roll that carries a 40,000 square foot distribution building carries the residential street behind it, with year built and roof cover on every house. A screen restricted to commercial stock is a choice, not a limit of the data.
The variation between counties is larger than people expect. One county publishes a clean parcel layer with a documented schema, refreshed quarterly, including roof material and improvement detail. Its neighbour publishes boundaries and an owner name, updates once a year, and codes land use in a scheme it has never documented. Both are the assessor.
- Owner of record commonly lags a transfer by a full assessment cycle, so the name in the file may not be the current owner.
- Year built refers to the original structure, not to the roof, the mechanical plant, or a substantial renovation.
- Land-use codes are assessment categories, not operating descriptions. A code says how a parcel is taxed, not what happens inside the building.
- Building area may be gross, heated, or footprint area depending on the county, and the difference matters if you are estimating usable roof.
- Multi-parcel campuses and condominium regimes routinely break a one-address, one-parcel assumption.
Building and electrical permit indexes
Permits are the closest thing to a dated event log for a building. A roofing permit with a 2021 issue date is a strong indication that a roof was worked on in 2021. An electrical or PV permit indicates a system was filed for and, in most jurisdictions, inspected.
What a permit proves is narrower than it looks. It proves an application was filed and, if the status field says so, that it passed inspection. It does not prove the work was completed as scoped, that it was done well, or that the equipment described is what sits on the roof today.
The absence of a permit proves even less. Some work is genuinely unpermitted. Some jurisdictions do not publish a searchable index at all. Some publish only the last few years. Some index by address string in a format that does not match the assessor's, which means the permit exists and your join missed it. Treat a missing permit as missing, never as evidence that nothing happened.
Elevation and terrain
USGS 3DEP gives bare-earth terrain, slope and aspect across a parcel. For ground-mount siting, drainage questions, and access grades this is the source that settles whether a site is buildable in the shape you assumed.
Coverage quality varies sharply. Where a county has been flown with lidar you get a high-resolution surface. Where it has not, you get a coarser DEM that will not resolve a small berm, a retention basin, or a change in grade across a single lot. Bare-earth also means the model has had structures and vegetation stripped out, so it does not describe what is standing on the site.
Hazard layers
FEMA's National Flood Hazard Layer gives the effective flood zone and floodway across a building footprint, with panel effective dates that differ county to county. NOAA storm event records give recorded hail and wind events, with dates, reported magnitudes and paths.
Both are regulatory or observational records rather than measurements at a property. A flood zone is a designation, and being outside a mapped zone is not the same as being outside flood risk. A storm report is a report of an event, not a report of what that event did to a particular roof.
Utility and tariff context
EIA and utility service-territory data establish which utility serves an address and the rate context around it. That is genuinely useful for deciding whether a territory is worth working at all, and for grouping properties by the programme rules that apply to them.
It does not establish anything about the meter. Actual consumption, demand profile, existing interconnection, available hosting capacity and queue position are not in these files. Those come from the customer, from the utility, or from an interconnection study.
How to audit a source list
Whether you are building this in house or buying it, the same questions apply to the finished report.
- Which specific sources were used in my counties, named, with capture or publication dates?
- Which of my counties are thin, and what is missing in each?
- For a given property, which source produced each field on the row?
- Where two sources disagreed, which one was preferred and why?
- Which fields are direct observations, which are public-record facts, which are derived estimates, and which are analyst judgments?
A firm that cannot answer those questions about your territory is asking you to trust a score. A firm that can is handing you something you can check, argue with, and have re-run. That difference is most of the value.
What none of this settles
Taken together these records can tell you which properties in a territory meet criteria you wrote down in advance, and where the evidence is weak. They cannot tell you structural or loading capacity, membrane or equipment condition, electrical capacity, interconnection feasibility, current ownership authority, energy consumption, or whether an owner will take the call.
Those are field, records, engineering and conversation questions. The point of the screen is to get you to them in the right order, on a shorter list, with the open checks already written down. Our methodology page names every programme above and states the same limits in the report itself.