Neighborhood Analysis Generator — From MLS Stats to a Buyer-Ready Brief

What it is

There is no established page for this job — search for a neighborhood analysis prompt today and you get CMA prompts, generic market-report prompts, and buyer-side research prompts scattered through listicles. The work itself, though, is something agents do constantly: a buyer asks "so what's that neighborhood actually like as a market?", a seller wants to know how their area is moving before a listing appointment, and a market-update video needs ninety seconds of accurate script.

The honest version of that job has a hard constraint: an AI has no live MLS access. It cannot look up a neighborhood. What it can do — brilliantly — is take the export you already pulled from your MLS and turn raw rows into a brief a human can use: ratios explained in plain English, two areas compared side by side, trends described without hype, and a script that says only what the numbers support. This generator builds the prompt that does exactly that, and its defining feature is what happens when the data runs out: claims get flagged [VERIFY], not smoothed over.

Who it's for

Buyer's agents preparing a neighborhood brief before a tour day; listing agents walking into an appointment with an area market story; agents who record monthly market-update videos or post market snapshots and are tired of writing the script from a spreadsheet; and any agent who has been handed a "just ask ChatGPT about the neighborhood" request and knows the answer has to come from real data — theirs.

How it works

1. Pull your stats first. Export or note the figures from your MLS for the area and period: median sale price, average days on market, list-to-sale price ratio, active listing count, pending count, and months of inventory or absorption rate if your MLS reports it. Same period for both areas if you're comparing.
2. Paste the stats into the field. Raw is fine — a copied table, a comma-separated list, even a paragraph. Note the period the figures cover; the prompt requires the output to cite it.
3. Add objective place facts only — named parks, transit stops, major employers, a waterfront, a farmers market — things that are checkable and neutral. Leave out anything about who lives there; the prompt will refuse it anyway (see Guardrails).
4. Generate, paste, and review the flags. The output is a comparison table, talking points, and a video script. Work through every [VERIFY] flag before the brief reaches a client: either source the item yourself or cut it.

What the output looks like

• Comparison table: your area (and a second area, if supplied) across every metric you provided, with the period labeled. Metrics you didn't supply appear as "not supplied" — never estimated.
• Plain-English talking points: each metric translated ("homes here are selling close to asking, and faster than last quarter" — only when your figures actually say that), written for a buyer or seller conversation.
• A 60–90 second video script: the three most defensible points from your data, in speaking order, with no claim the table doesn't support.
• A [VERIFY] list: every statement the AI wanted to make that your data doesn't back — school ratings, safety impressions, "up-and-coming" characterizations — quarantined in one place for you to source or delete.

THE PROMPT TEMPLATE

You are a real estate market analyst preparing a neighborhood brief
for a licensed agent. You have NO live data access. Everything you
write must be derived from the statistics I paste below. If a claim
cannot be derived from them, it does not go in the brief — it goes
in the [VERIFY] list.

AREA: [Maplewood / ZIP 98103]
COMPARISON AREA (optional): [Greenwood / ZIP 98103 — north]
PERIOD COVERED BY ALL FIGURES: [Sep 1 – Sep 30, 2026]

MY MLS STATISTICS (the complete allowed fact set):
- Median sale price: [$672,000]
- Average days on market: [21]
- List-to-sale price ratio: [99.1%]
- Active listings at period end: [34]
- Homes sold in period: [41]
- Months of inventory: [0.83]
- (Comparison area, same period: median [$715,000], DOM [26],
  list-to-sale [98.4%], active [29], sold [35], inventory [0.83→
  use supplied figure])

OBJECTIVE PLACE FACTS (neutral, checkable — amenities and places only):
- [Green Lake Park and trail loop at the south edge]
- [Route 5 bus corridor on Greenwood Ave]
- [North Seattle College approx. 1.5 miles east]

TASK — produce three deliverables:
1. COMPARISON TABLE: both areas side by side across every supplied
   metric, period labeled at the top. Any metric not supplied for
   an area reads "not supplied."
2. TALKING POINTS: 5–7 bullets in plain English for a client
   conversation. Translate each metric honestly (what a list-to-sale
   ratio near 99% means, what under one month of inventory means).
   Balanced: include whatever the data says that cuts against a
   rosy reading, too.
3. VIDEO SCRIPT: 60–90 seconds spoken, built only from the three
   strongest data-backed points, conversational, no hype words.

STANDING RULES — NEGATIVE CONSTRAINTS (never break these):
- Use ONLY the supplied statistics and place facts. No live data,
  no remembered data, no estimates, no "typically" or "usually"
  claims about this area.
- Anything you believe but cannot derive from the supplied data
  goes in a final "[VERIFY] — not sourced from supplied data" list.
  This includes ALL school ratings or quality claims, ALL safety
  or crime characterizations, and ALL demographic descriptions.
- Describe amenities and places, never residents: no statements
  about who lives in either area, what "kind of people" an area
  suits, or which area is "better for" any group (Fair Housing).
- No price predictions, no timing advice ("buy now before prices
  rise"), no investment guarantees. Report the period's data;
  do not forecast.
- Date-stamp every claim to the period covered. Last period's
  figures are never described as current conditions.

OUTPUT FORMAT:
- Table, then Talking points, then Video script, then the [VERIFY]
  list, then "Assumptions made: none" — or list them.

Best model for this: Gemini 4 Argon — it handles a long pasted stats
table comfortably and is reliable at producing clean comparison
tables and structured summaries from supplied data.

ONE WORKED EXAMPLE

Inputs filled in: the Maplewood figures in the template above (median $672,000, 21 days on market, 99.1% list-to-sale, 0.83 months of inventory, period Sep 2026), Greenwood as the comparison area, and three objective place facts. Purpose: buyer brief.

The filled prompt is the template with those values in place.

Excerpt of the output it produces:

The example is the feature: the analysis is genuinely useful and its two weakest impulses — a school claim and a demographic claim — are caught, labeled, and quarantined where the agent can deal with them deliberately.

GUARDRAILS — steering, schools, safety, and the [VERIFY] rule

Neighborhood content is where Fair Housing risk concentrates, because "describing an area" slides almost imperceptibly into describing its people — and steering a buyer toward or away from an area based on who lives there is a Fair Housing violation whether a human or an AI phrases it. This page is built around three hard rules.

Describe amenities and places, never residents. Parks, transit, employers, a lake, a commercial strip: fine. "Young," "upscale," "diverse," "family-oriented," "up-and-coming" as people-descriptions: out — both in what you type into the fields and in what the prompt allows out. If a client asks you directly which area is "better for people like us," the compliant move is the one this tool models: hand them objective data and official sources, and let them decide.

No school ratings and no safety claims from an AI. School quality characterizations and crime impressions are the two claims clients most want and an AI is least able to source. The prompt quarantines every one of them into the [VERIFY] list automatically. The professional workflow: point buyers to the school district's own enrollment and boundary information and to official public safety data sources, and let them evaluate. Never let an AI's adjective do that job in a client-facing brief.

Supplied data only, period stamped, no forecasts. The prompt forbids remembered or assumed statistics, requires every claim to trace to your pasted figures, and bans price predictions and buy-now timing advice outright. A brief that says "in September, homes here sold in a median of 21 days" is defensible forever. A brief that says "this neighborhood is hot and prices will keep climbing" is a liability the day it is written.

The Standing Rules recap: supplied statistics and place facts only; everything else to the [VERIFY] list; amenities and places, never residents; no school or safety claims; no predictions or guarantees; "Assumptions made" declared at the end. The agent still reviews the final brief — the tool's job is to make the risky parts impossible to miss.

INTERNAL LINKS

• Real Estate hub: https://promptstaging.site/generators/real-estate/
• Real Estate guide (full workflow article): https://promptstaging.site/article-real-estate.html
• Listing Description Generator: https://promptstaging.site/generators/real-estate/listing-description.html
• Follow-Up Email Sequence Generator: https://promptstaging.site/generators/real-estate/follow-up-email-sequence.html
• Objection Handler: https://promptstaging.site/generators/real-estate/objection-handler.html
• Homepage Prompt Generator: https://promptstaging.site/#tool

---