Cutting a 100-candidate rooftop list to 10 site visits with height data

A search ring in a mid-density suburb can throw forty, sixty, a hundred rooftop candidates at you before you've even opened the parcel database. Somebody still has to get that down to a shortlist a crew can walk in a week. The usual method is a planner scrolling street view, checking old permit PDFs, and guessing at storey count from a photo taken five years ago. That guess is the thing that gets a site cut from the visit list for the wrong reason, or kept on it when it never had a shot.
What desk screening can rule out before anyone drives out
Desk screening earns its keep by stopping you from paying mileage and a half-day visit fee on buildings that fail on height alone, long before anyone picks a winning roof. If your RF design needs 15 m of clearance above the surrounding roofline and a candidate sits at four storeys in a neighborhood of two-storey rowhouses, that's useful information sitting in the data before anyone touches a tape measure. The same goes the other way: a building that looks promising on the parcel map but turns out to be a single-storey retail box with a false parapet reads tall from the street and short from above.
Zoning records help with setbacks and use class, but they go stale fast and they were never built to tell you floor count. A building permitted as three storeys in 2009 might have added a rooftop mechanical penthouse since, or lost a floor to fire damage and rebuild. Planning records show what was approved at permit time, which drifts further from what's standing today the longer the record sits unreviewed.
Where a height layer narrows the list
This is the part that turns a hundred candidates into ten. Run the candidate list against a per-building height and estimated storey value, and you can sort by clearance margin before you write a single visit itinerary. Anything that fails your minimum height threshold by a wide margin drops off immediately. Anything that clears it comfortably and sits in a gap in your existing grid moves to the top. The middle tier, the ones within a storey or two of your cutoff, is where a site visit earns its cost: those are the candidates where the field numbers need to override the desk numbers.
That's a different use of the data than most telecom planners default to. It's not about getting an exact antenna mount elevation from a desk tool, because you still want a structural engineer and a site visit for that. It's about triage: using a current, building-by-building height read to decide which ten (or fifteen, or twenty) addresses deserve the field team's time this quarter, out of a list where most candidates were never going to clear your RF requirement in the first place.
The site storeyestimation.com built around exactly this gap turns VHR satellite or aerial imagery into a per-building height and storey estimate, delivered as a CSV or a layer attribute you can join to your candidate list directly, refreshed yearly so you're not screening against a seven-year-old survey.
Ordering the shortlist, not just filtering it
Once the obvious fails are cut, the remaining list still needs an order. Pair the height read with whatever RF priority score your team already uses, lease cost estimates, landlord responsiveness history, and you get a visit sequence instead of a flat list of ten addresses. Send the crew to the candidates where height, RF fit, and lease terms all line up first. Hold the marginal-clearance buildings for a second pass once the first batch either lands a lease or falls through on something the desk data couldn't have caught, like a landlord who won't return a call or a roof access easement that doesn't exist on paper.
None of this replaces the site visit. A planner still needs boots on the roof to confirm structural load, actual parapet height, and cable run distance. What a height layer does is stop you spending that visit budget on buildings that were never tall enough to begin with.
If your next search ring is sitting at eighty or a hundred candidates and the visit budget covers ten, see what Storey Estimation's height and storey data can cut before the crew ever loads the van.