Update to Advan Research REI

{
  "name": "trade-area-analysis",
  "description": "Workflow for defining and analyzing the trade area of a retail property using Advan REI mobile visitor origin data. Use this skill whenever a user asks where a property's customers are coming from, how far people travel to visit a location, what the catchment area looks like, or anything about the geographic reach of a property. Also trigger when a user asks about trade area as part of a site selection, leasing, or competitive analysis — even if they don't use the phrase \"trade area\" explicitly. Phrases like \"where are visitors coming from\", \"how far do customers drive\", \"what's the catchment\", \"customer origin\", or \"who is in the surrounding area\" should all trigger this skill. Works equally well as a standalone analysis or as a follow-on to foot traffic data.\n",
  "included_files": [],
  "skill_md_contents": "---\nname: trade-area-analysis\ndescription: >\n  Workflow for defining and analyzing the trade area of a retail property using\n  Advan REI mobile visitor origin data. Use this skill whenever a user asks\n  where a property's customers are coming from, how far people travel to visit\n  a location, what the catchment area looks like, or anything about the\n  geographic reach of a property. Also trigger when a user asks about trade\n  area as part of a site selection, leasing, or competitive analysis — even if\n  they don't use the phrase \"trade area\" explicitly. Phrases like \"where are\n  visitors coming from\", \"how far do customers drive\", \"what's the catchment\",\n  \"customer origin\", or \"who is in the surrounding area\" should all trigger\n  this skill. Works equally well as a standalone analysis or as a follow-on\n  to foot traffic data.\n---\n\n# Trade Area Analysis Workflow\n\nThe trade area analysis answers a different question than foot traffic: not\nhow many people visit, but *where they come from*. The output is geographic —\na picture of the real catchment area built from actual visitor origin data,\nnot assumed drive-time rings.\n\nThe workflow has four stages: resolve the property, pull the trade area and\ncharacterize its size, layer in a demographic snapshot, and synthesize with\nreal-world context.\n\n---\n\n## Stage 1: Resolve the Property\n\nSame as foot traffic. If the user has already provided or confirmed an ALI,\nlocation name, and company name — or if these were established earlier in the\nconversation — carry them forward and proceed.\n\nIf the property is ambiguous, use `search` to find it, present the top\nresults, and confirm before pulling data.\n\n---\n\n## Stage 2: Pull the Trade Area\n\nCall `trade-area` with the resolved property identifiers.\n\n**Parameters:** Determine the threshold in this order:\n\n1. **User specified it in the conversation** (e.g., \"show me the 50% trade\n   area\", \"use the 80% catchment\") → use that, regardless of app settings.\n2. **User is in the REI app and hasn't specified** → use whatever threshold\n   they have configured in their trade area settings.\n3. **No app, no specification** → default to the closest 70% of visitors.\n\n**Date range:** Use the period the user has selected in the UI, or whatever\nthey've specified. Only ask if it's genuinely unclear.\n\n### Characterize the size\n\nThe trade area is defined by where visitors actually originate, not by an\nartificial ring — but users need a tangible sense of scale. From the returned\ndata, derive and report the approximate size in one of these terms, whichever\nis most natural given the geography:\n\n- **Radius:** \"The trade area extends roughly X miles from the property\"\n- **Drive time:** \"Most visitors are within a Y-minute drive\"\n- **Both:** Use both when the shape is notably asymmetric or the market\n  warrants it (e.g., a coastal or highway-adjacent location where drive time\n  and radius diverge meaningfully)\n\nIf the trade area is unusually compact or unusually large relative to what\nyou'd expect for the property type, flag it — it may reflect a dense urban\nsetting, strong competition nearby, or a destination-oriented tenant mix.\n\n---\n\n## Stage 3: Demographic Snapshot\n\nCall `demography` for the trade area using the same property and date range.\n\nBy default, surface a concise top-line snapshot — enough to characterize\nwho's in the catchment without overwhelming the response. Focus on the\nsignals most relevant to a real estate or retail professional:\n\n- Median household income (and whether it skews upper/lower relative to\n  the metro)\n- Dominant age cohort(s)\n- One or two standout lifestyle or psychographic segments if they're\n  meaningfully concentrated\n\nEnd the snapshot with a natural offer to go deeper:\n\n> \"Want me to break this down further by income band, age distribution, or\n> lifestyle segment?\"\n\nDon't pull `demography-categories` unless the user asks for the full\nbreakdown — it's there for drill-down, not the default view.\n\n---\n\n## Stage 4: Synthesize and Present\n\n### Search for real-world context\n\nBefore writing the narrative, do a targeted web search to find anything that\nhelps explain or enrich what the data is showing about the catchment. Useful\nangles include:\n\n- Economic character of the trade area (affluent suburb, dense urban core,\n  working-class exurb, etc.)\n- Major employers or demand generators within the catchment\n- Population growth or demographic shifts in the area\n- Infrastructure or access factors that shape the catchment (highways,\n  transit, topographic barriers)\n- Competitive landscape — are there strong competitors drawing from the\n  same geography?\n\nAs with foot traffic: if the search turns up something that connects, weave\nit in. If it doesn't, move on. The goal is context that helps the user\ninterpret the trade area, not filler.\n\n### Lead with the headline\n\nOne sentence that captures the essential story — size, character, and\nanything notable.\n\n> \"Riverside Commons draws the bulk of its visitors from a roughly 4-mile\n> radius, a trade area anchored by the upper-middle-income neighborhoods of\n> Maplewood and Crestview, with a median household income around $95K.\"\n\n### Trade area summary\n\nAfter the headline, a short structured summary:\n\n| | |\n|---|---|\n| Trade area threshold | 70% of visitors (or whatever was used) |\n| Approximate size | X miles / Y-minute drive |\n| Median HHI | $X |\n| Dominant age cohort | X–Y |\n| Key lifestyle segments | [top 1–2] |\n\nKeep this tight. The table is a reference, not the analysis.\n\n### Narrative\n\nTwo to three sentences connecting the geographic and demographic picture to\nsomething actionable. What does this trade area mean for the property? Think\nabout:\n\n- Whether the catchment is concentrated or dispersed (and what that implies\n  for co-tenancy or marketing reach)\n- Whether the demographic profile aligns with the current tenant mix — or\n  suggests an opportunity\n- Any geographic constraint or advantage worth flagging (a highway that\n  extends reach to the east, a river that cuts it off to the west)\n- Anything from the web search that adds color\n\n### Cite data vintage\n\n> *Trade area and demographics based on [date range], via Advan's mobile\n> device panel. Trade area defined as the closest [X]% of visitors by origin\n> (per [user's REI app settings / default threshold].)*\n\n---\n\n## Handling Gaps and Anomalies\n\n**No trade area data returned:** Say so plainly and offer to escalate or\ncheck a comparable.\n\n**Unusually small or large catchment:** Flag it and offer a hypothesis if\none is obvious from context (dense competition, destination anchor, barrier\ngeography). Don't just report it as a fact without comment.\n\n**Demographic data sparse or incomplete:** Present what's available, note the\ngap, don't fill it with estimates.\n\n---\n\n## Follow-Ups\n\nIf there's a natural next direction given what the data showed, mention it.\nThe most common ones that arise organically from trade area work are\ncross-shopping (who else in the catchment are these visitors going to?),\nvoid analysis (what tenant categories are underrepresented given this\ndemographic profile?), and competitive context (how does this catchment\ncompare to a neighboring center?). But let the conversation lead — don't\nrecite options.\n"
}