costinflation/mini-split-ac-prices-raw-dataset-2026

25,761 raw U.S. mini-split AC price observations across 12 ZIP markets and 29 days.

Mini-Split AC Prices Raw Dataset (2026)

Analyze 25,761 unaggregated product-level listed retail prices for complete mini-split air-conditioning systems across 12 U.S. ZIP markets from July 13 through August 10, 2026. The single analysis-ready CSV preserves titles, dates, geography, package quantities, listed prices, and a source-neutral comparable-price field.

What “raw” means here: unaggregated product-level observations after scope and safety filtering. The file includes derived normalization fields; it is not untouched acquisition data.

Dataset at a glance

Field	Coverage
Rows	25,761
Columns	16
Observation window	July 13–August 10, 2026
Dates / ZIP markets	29 / 12
Market-days	348 of 348
Distinct product titles	431
Comparable-price rows	25,761 (100.0%)
Rows with unresolved comparison	0 (0.0%)
Comparison target	1 complete system
Format / currency	CSV / USD
Expected updates	None; fixed research snapshot

A result worth investigating

Listed-price P90 was 2.5× the P10. Across one-unit observations, listed prices ranged from a P10 of $386.99 to a P90 of $949.99, a 2.5× spread. This is a title-level retail distribution, not a claim of product, clinical, installation, or performance equivalence.

What one row represents

One row is one quality-filtered product-title, ZIP-market, date, package-quantity, and listed-price observation. These are listed retail prices—not sales, orders, revenue, demand, inventory, market share, or consumer behavior. Multiple distinct offers can share a title, ZIP, and date; exact duplicate public rows have been removed.

Comparable-price method

The fixed comparison target is 1 complete system. This is one-unit listed-price comparability. It is not installed cost, efficiency, capacity, or performance equivalence. The original package price and quantity remain in every row, so users can audit or replace the provided comparison.

Columns

Column	Type	Meaning
series_id	string	Stable source-neutral dataset series key
series_title	string	Human-readable series title
canonical_url	string	Related CostInflation category page
geography_type	string	Geography level; always postal_code
geography_id	string	Five-digit U.S. ZIP code; load as text
geography_label	string	Human-readable ZIP-market label
observed_date	date	Observation date in YYYY-MM-DD
product_name	string	Full observed title; not a stable product ID
quantity_value	number	Resolved package quantity; 0 if unresolved
quantity_unit	string	Resolved package unit or unknown
quantity_name	string	Human-readable package quantity
price_amount	number	Listed package or unit price
currency_code	string	Currency; always USD
normalized_price_amount	nullable number	Listed price scaled to the dataset comparison target
normalized_quantity_value	number	Numeric comparison target
normalized_quantity_unit	string	Unit for the comparison target

Project ideas

• Explore listed-price dispersion across system titles
• Track matched mini-split titles through time
• Study cross-market listed-price parity
• Build an HVAC retail dashboard
• Run EDA in Python, R, SQL, Excel, Tableau, or Power BI
• Benchmark missing-value handling without discarding listed-price observations

Quick start with Pandas

import pandas as pd

df = pd.read_csv(
    "costinflation-mini-split-ac-retail-prices-raw-2026-07-13-to-2026-08-10.csv",
    dtype={"geography_id": "string"},
)
comparable = df.dropna(subset=["normalized_price_amount"]).copy()
print(comparable["normalized_price_amount"].describe())

Quality and limitations

• All 348 expected date × ZIP-market combinations are present.
• Exact duplicate public rows: 0. The preparation pass removed 0 byte-identical input repeats.
• Scope review excluded 1 definite out-of-scope or unsafe titles representing 8 input rows.
• Product titles are descriptive text, not stable public product identifiers.
• ZIP labels describe selected markets, not citywide estimates; the panel is not nationally representative.
• Availability and title wording can change unmatched aggregates through assortment change.
• Shipping, tax, redeemed promotions, purchases, and product performance are outside the dataset.

License

Released under CC0 1.0 Universal for unrestricted reuse. Attribution is not required, but citation helps others find the release.

CostInflation Team. (2026). Mini-Split AC Prices Raw Dataset (2026). CC0 1.0 Universal.

Resources

• Mini-split AC price history
• The matching Hugging Face or Kaggle release will be linked after both destinations are verified.

Released by CostInflation Team.