Home Depot Investor Relations – SEC Filings & Financials

The investor-facing record of the largest home improvement retailer on earth (NYSE: HD, SIC 5211, fiscal year ending late January or early February), opened up as flat records. The surface spans quarterly earnings releases, annual reports, proxy statements, current forms and the complete SEC filing trail; the currently listed key reports run to the 2026 Proxy Statement, the 2025 Annual Report, the FY25 10-K and the 1Q26 10-Q.

Scale and substance first. Fiscal 2025 sales came in at $164.7 billion against $159.5 billion the prior year, adjusted operating margin printed 13.1 percent versus 13.8 percent, and adjusted ROIC landed at 25.7 percent. Beneath the headlines, the filings carry machine-readable XBRL tagging: every reported concept arrives as a typed number with its period attached, which is what turns a stack of several-megabyte documents into a queryable series. Recent forms alone enumerate 97 8-Ks and 25 10-Qs alongside multiple 10-Ks, with history reaching back through the company's IPO era.

Datadory packages that record as one documented deliverable: tagged facts, filing index and headline metrics normalized into the field dictionary below. It scores 9 on Datadory's ten-point quality rubric with field definitions verified against captured samples. [Get a sample of this dataset](#request) and judge the columns, not the promise.

One record per tagged fact, exactly as it lands in your warehouse. Both rows are genuine captured values, not illustrations:

concept : RevenueFromContractWithCustomerExcludingAssessedTax
start   : 2025-02-03      end     : 2026-02-01
val     : 164683000000
form    : 10-K            filed   : 2026-03-18
fy      : 2025            fp      : FY

concept : RevenueFromContractWithCustomerExcludingAssessedTax
start   : 2026-02-02      end     : 2026-05-03
val     : 41765000000
form    : 10-Q            filed   : 2026-05-27
fy      : 2026            fp      : Q1

Read it as a time-series kit. Each fact carries its own period boundaries, so quarter-over-quarter growth is arithmetic rather than reconstruction; the form type tells you whether a number is audited annual reporting or an unaudited interim print; and the filing date lets you reproduce any point-in-time view of what the market knew when. The second row shows why the shape matters: $41.77 billion for the first quarter of fiscal 2026, sitting beside the full-year figure in the same schema, joins into a continuous series without remapping.

Eight documented fields anchor the dictionary, verified against captured samples during research and held stable across captures. Surfaces whose location shifts between filings - comparable-sales commentary, store-count tables, department-level sales - are held under additional fields on request rather than promised in every row, because promising a column whose position moves year to year is how pipelines break.

Geography - consolidated reporting for the US-headquartered company, spanning United States operations plus Canada, Mexico and other international footprints rolled into one set of statements. One entity, one consolidation boundary, no guesswork about which subsidiary a number belongs to.

Temporal - filing history reaching back through the company's IPO era, which makes this the longest continuous quantitative record of home-improvement demand available anywhere: decades of point-in-time fundamentals rather than a snapshot. Recent history includes the FY25 10-K filed 2026-03-18 and the 1Q26 10-Q filed 2026-05-27.

Granularity - one record per tagged fact and one per indexed filing, at consolidated quarterly-and-annual depth. Segment and department detail rides inside the filing documents rather than flattened into the fact table, which keeps the core contract small enough to learn once and deep enough to reward a second look.

API, files, or your warehouse. Daily, weekly, or hourly.

Daily suits event-driven work, where an 8-K landing between reports moves a thesis before most desks have parsed it. Weekly suits fundamentals tracking, where the quarterly rhythm matters more than intraday noise. Around each report date a burst capture keeps the new 10-K or 10-Q flowing into your tables the day it lands. Whichever cadence you pick, join keys stay stable: the accession number back to the filing index, the fiscal-year and period labels onto your own calendar, concept names onto your chart of accounts.

• Fundamentals pipelines and ML features - every value arrives typed, dated and attributable to a filing, ready for model training; mapped to the ML model training use case.
• Point-in-time quant research - the filing-date trail builds training sets without look-ahead, feeding quant backtesting.
• Market sizing and benchmarking - one company's multi-decade record anchors home-improvement economics instead of assembled press clippings; see the market sizing use case.
• Competitive positioning - strategy and category-mix disclosures track where the category-defining retailer is heading; mapped to competitor tracking.
• Demand modeling - reported sales paired with federal retail aggregates give consumer-demand features at both company and category level; see demand forecasting.
• Citation-grade research - journalists, academics and students quote public-record filings with a paper trail behind every figure.

Data scientists and ML engineers get typed, dated, attributable facts - the difference between a fundamentals pipeline and a parsing project. Investors and quants model the home-improvement leader from audited statements with point-in-time integrity built in. Market researchers and consultants benchmark category economics from one multi-decade record. Competitive intelligence teams read the strategy disclosures of the retailer that defines the category. Journalists, academics and students get citation-grade sourcing behind every number. Routes for data scientists, competitive intel product teams and investors and quants all run through this dataset first.

Three things worth knowing upfront. First, comparable-sales commentary appears as prose in earnings releases and management discussion rather than as a tagged column, so if your model needs comp-store growth as a number, name that when requesting the sample and the extraction gets scoped explicitly. Second, store counts by state and sales-by-department tables live inside the 10-K documents themselves, and their page positions shift between fiscal years - quoted as scoped work, never discovered mid-project. Third, this record is deliberately narrow: one company's economics, not the whole home-improvement market. Pair it with the Census monthly retail aggregates in the same home furnishings hub when you need the category-level benchmark beside the company-level truth.