The single best ZIP-level home-value and rent series a free source publishes. A Zillow skill teaches Claude to read ZHVI against the long-run state / metro trend and to use ZORI as the rent-benchmark it already is for every real investor's underwriting spreadsheet.

Reference US home values, rents, listings flow, sales velocity, price cuts, and affordability from Zillow (Zillow Research public CSVs) for real estate investment decisions. Use this skill whenever a user asks about ZHVI, Zillow Home Value Index, ZHVF, ZORI, ZORDI, ZORF, PITI, for-sale inventory, newly pending listings, sale-to-list ratio, days to pending, price-cut share, Zillow Market Heat Index, new construction prices, years to save, ZIP-level home values, metro rent comps, or which Zillow Research cuts inform a residential RE decision. Maps 33 Zillow headlines + 119 variants across 11 categories (home_values, mortgage_carry, rents, inventory_listings, prices_sales, negotiation, velocity, price_cuts, market_heat, new_construction, affordability) to investor use-cases. Progressive disclosure — decision playbooks and the full 33-headline reference live in `references/`.

Zillow for Real Estate

What is Zillow

Zillow Research is Zillow's public data program — zillow.com/research/data. Free, no signup, ~150 monthly CSVs covering home values, forecasts, rents, listings flow, sales, days-on-market, price cuts, market heat, new construction, and affordability. For US residential real estate, it's the authoritative single stop for ZIP-level home values (ZHVI), metro-level rent benchmarks (ZORI), and the listings-side velocity reads no government source publishes monthly.

Zillow Research goes ZIP-deep on home values and metro-deep on most listings/sales/affordability cuts. For below-ZIP grain (tract, block-group) on demographics and tenure, pair with Census ACS — Zillow doesn't publish at that resolution.

Bundled resources

Progressive disclosure rule: start with this file and the category table below. Read references/metrics.md only when a specific series, variant, or CSV URL is actually needed. Read the YAML only when the markdown is insufficient.

Getting an API key

1. No API key required. Zillow Research publishes static CSVs at files.zillowstatic.com/research/public_csvs/... — pd.read_csv(url) is the entire client.
2. No env var to set, no signup, no rate limit published. Be a good neighbor: cache locally, don't hammer the same URL on a loop.
3. Refresh cadence is monthly for most series, weekly for a handful of variants (look for _week.csv vs _month.csv in the URL).
4. Terms of use: research/non-commercial OK, commercial redistribution requires permission. Cite Zillow Research when you publish.

Minimal Python example

import pandas as pd
 
# ZHVI All Homes (Smoothed, SA, Mid-Tier) — Tampa MSA (RegionID 395148)
url = (
    "https://files.zillowstatic.com/research/public_csvs/zhvi/"
    "Metro_zhvi_uc_sfrcondo_tier_0.33_0.67_sm_sa_month.csv"
)
df = pd.read_csv(url)
meta = ["RegionID", "SizeRank", "RegionName", "RegionType", "StateName"]
tampa = (
    df[df["RegionID"] == 395148]
    .melt(id_vars=[c for c in df.columns if c in meta],
          var_name="date", value_name="zhvi")
    .dropna(subset=["zhvi"])
)
print(tampa[["date", "zhvi"]].tail(12))

Swap the URL for any CSV from references/metrics.md (or the csv_url / variants[].csv_url fields in the YAML). Featured geography across the catalog is Tampa-St. Petersburg-Clearwater MSA (RegionID 395148), benchmarked against Florida (RegionID 14) + United States (RegionID 102001).

The 11 categories

Total: 33 headlines + 119 variants = 152 cuts. See references/metrics.md for the full list with dataset_keys, CSV URLs, and the per-headline "Also available as..." variant lists.

Decision playbooks

Concrete "the user asked X — here's the Zillow series set to pull." Use these as starting points; extend when the question demands it. If a playbook names a series, references/metrics.md has the verbatim rationale.

"What's a typical home worth here, and where's it headed?" → zillow_zhvi, zillow_zhvf, zillow_median_sale_price ZHVI is the defensible "typical home value." ZHVF is the directional forecast. Median sale price is the noisier transaction read — never use it without ZHVI alongside.

"Is this market a buyer's or seller's right now?" → zillow_market_heat_index, zillow_sale_to_list_ratio, zillow_share_sold_above_below_list, zillow_share_price_cut Heat index is the composite. Sale-to-list and above-list share read negotiation directly. Price-cut share is the early-warning that sellers are blinking before the ratios move.

"How fast is supply turning, and is it tightening or loosening?" → zillow_for_sale_inventory, zillow_new_listings, zillow_newly_pending_listings, zillow_days_to_pending Inventory is the level. New listings vs newly pending tells you whether supply is growing because nothing sells or because sellers are listing. Days to pending is the velocity confirm.

"Can a typical buyer afford anything in this metro?" → zillow_new_homeowner_affordability, zillow_new_homeowner_income_needed, zillow_affordable_home_price, zillow_years_to_save Affordability index is the share-of-income read against Zillow's 30% threshold. Income needed and affordable price work the same math from opposite ends. Years to save is the down-payment angle.

"What's a fair rent ask, and where's rent demand going?" → zillow_zori, zillow_zordi, zillow_zorf, zillow_new_renter_affordability ZORI is the rent benchmark. ZORDI leads ZORI by months — demand climbs first, rent follows. ZORF is the national forecast. Renter affordability checks whether your ask is breaching the 30% threshold for local renters.

"Is the new-construction market diverging from resale here?" → zillow_new_construction_median_sale_price, zillow_new_construction_median_price_per_sqft, zillow_new_construction_sales_count, zillow_median_sale_price Builder vs resale median tells you the new-build premium (or discount). Price-per-sqft normalizes for builders shrinking product. Sales count says whether builders are actually moving inventory.

Gotchas

• Smoothed vs raw, seasonally-adjusted vs not. Most headline series default to smoothed + SA (_sm_sa_month.csv); raw and non-SA cuts ship as variants. Don't compare a smoothed-SA series against a raw one without normalizing — the smoothing strips real moves alongside noise.
• CSV columns are dates, not values. Every Zillow CSV is wide-format: meta columns (RegionID, SizeRank, RegionName, RegionType, StateName) plus one column per month/week. Always melt() after filtering on RegionID.
• Geo levels vary per series. ZHVI ships at national/state/metro/county/city/ZIP/neighborhood. Many sales/listings cuts are national + metro only. Check geo_levels in the YAML before assuming ZIP coverage.
• Tier suffixes encode price band. tier_0.33_0.67 = mid-tier (33rd–67th percentile, the headline). tier_0.0_0.33 = bottom tier. tier_0.67_1.0 = top tier. Bedroom counts come in via bdrmcnt_N. Mix tier and bedroom in URLs — don't hand-edit, lift from the variant list.
• ZHVF is a growth forecast, not a level. ZHVF columns are MoM%, QoQ%, YoY% — apply to current ZHVI to project. Do not chart ZHVF as a price level.
• Nowcast cuts revise. Sales count, median sale price, mean sale price, and total transaction value publish "nowcast" variants that get revised as off-market data lands. Pull the latest CSV every time; don't cache the trailing month's number.
• No published rate limit, but be polite. Zillow doesn't enforce per-IP throttling on the public CSVs, but they're hosted on a static CDN. Cache locally; refresh monthly aligns with the publishing cadence.
• Forecast is national-only for rent. ZORF (rent forecast) ships at national grain. For metro rent calls, pair ZORI's trailing trend with your own top-down — Zillow doesn't publish a metro-level ZORF.

Source of truth

This skill is derived from TiRE's canonical catalog at agents/data/zillow-datasets.yaml. The YAML is hand-curated; every entry has a why_it_matters rationale that feeds both this skill and TiRE's research agents.

When the upstream catalog changes, regenerate references/metrics.md and references/zillow-datasets.yaml. v1 is hand-curated; build script deferred until drift is observed.