What this is. A repeatable flow that adds five local-interest layers to a buyer tour map (e.g. the Thode & Cannon Heights tour). The point isn't just to list shops — it's a method: every category has a generic search recipe keyed to one input — {neighborhood, centroid, radius, client profile} — so the same flow regenerates the layer for River Oaks, Memorial, or any city node with zero rework. Grounded examples below are from The Heights, Houston; swap the inputs and the recipe finds the equivalents anywhere.
The 5 Local-Interest Categories + Generic Search Method
Each recipe is written to run against a Places search (Google Places / Maps) around the tour centroid. Inputs are always the same four variables, so the recipe is area-agnostic.
🛒 1 · Grocery anchors
"Where do I do my weekly shop?"
Generic recipe: Text-search a fixed named-brand anchor list near {centroid} within {radius}: H-E-B, Whole Foods Market, Kroger (+ tier-2: Trader Joe's, Costco, Aldi, Sprouts). Keep the nearest of each brand; record distance + drive time from the tour.
Why named brands: they're consistent national/regional anchors, so the recipe is identical in every market — only the nearest locations change.
🔧 2 · Hardware anchors
"Where do I get home & yard stuff?"
Generic recipe: Text-search named-brand list near {centroid}: The Home Depot, Lowe's (+ tier-2: Ace Hardware, Tractor Supply). Keep nearest of each; distance + drive time.
🏬 3 · Shopping centers (curated destinations)
"Where's the walkable retail + dining destination?"
Generic recipe: Places search type=shopping_mall plus keyword queries "mixed-use development" / "shopping center" / "market" / "mercantile" near {centroid}. Keep results with ≥6 tenant POIs and rating ≥4.3; rank by tenant count × rating.
Heights examples found: Heights Mercantile, M-K-T (adaptive-reuse retail/dining).
📍 4 · Store density (is it walkable & alive?)
"Is this a lively, walkable pocket — not a strip-mall void?"
Generic recipe: Places Nearby type=store + type=restaurant within the walk radius (≤0.5 mi). Count POIs per 0.25-mi grid → density score. Separately, detect named retail corridors by searching "{main street} shopping". Output a walkability score + the live corridor names.
Heights examples found: 19th Street shopping district, The Houston Farmers Market — high-density walkable corridors.
⭐ 5 · Local favorites (local vibe, not chain)
"The independent, beloved spots — the character the buyer asked for."
Generic recipe: Places Nearby in high-signal categories (cafe, bakery, bar, specialty food, bookstore, roaster) near {centroid}, then filter: rating ≥4.5, reviews ≥300, and NOT on the national-chain denylist (Starbucks, Chili's, etc.). Rank by rating × review count.
Heights examples found: Houston Dairymaids (artisan cheese), Lei Low (tiki bar). The chain-denylist is what makes "local vibe ≠ chain" a reproducible filter, not a judgment call.
The Build Flow — 6 Stages
0
Inputs (Spec-Lock)
Lock the four variables: {neighborhood + city}, {centroid} (or the property address set), {radius / walk-threshold}, and the {client lifestyle profile} (from email/CRM — which categories matter most, e.g. "café ≤4 blocks, no chains"). parameterized
1
Run the 5 recipes
Execute each category recipe above around the centroid. This is the area-agnostic core — identical code, different inputs.
2
Gather & geocode
De-dupe, geocode every POI, compute distance + walk/drive time from the nearest tour stop, tag each: category · distance · walkable(Y/N) · client-priority match.
3
Score & curate
Rank within each category; keep top-K (≈3 grocery · 2 hardware · 3 shopping · top density corridors · 5 local favorites). Build a per-home "amenity fit" line (nearest grocery/hardware/coffee + walk score) and one neighborhood-level summary.
4
Render onto the tour map
Add three branded sections to the tour page — 🛒 Daily-life anchors (grocery + hardware) · 🏬 Shopping & local scene (shopping centers + density corridors, embedded map) · ⭐ Local favorites (independents, map-pinned). Each home card gets an amenities mini-line. Navy/gold/cream, mobile-first, noindex (private client).
5
Verify & ship
curl the rendered page (200 + sections present + real values, not placeholders); every POI link opens the correct Maps place; deploy to the live alias + verify; log to the Content tracker / Search hierarchy.
Replication — "how would you search these categories in another area?"
That's the whole design: Stage 1's five recipes are the answer. They never change — only {neighborhood, centroid, radius, client profile} change. Point the same flow at River Oaks, Memorial, or a Los-Angeles / New-York node and it regenerates the layer there.
This becomes a registered, parameterized engine — local_interest_layer(neighborhood, centroid, radius, client_profile) — so it's a tool the network reuses, not a one-off page. It plugs in as a component of the Showing-Tour SOP (and feeds the live mobile "Conduct" companion — coordinated with Vishal's tour build, not duplicated).