Datadory notebook
Consumer Electronics Demand Forecasting Dataset: What to Train On
Consumer electronics demand forecasting datasets start with the U.S. Census Monthly Retail Trade Survey: monthly NAICS 443 sales and matching inventories from January 1992, free public-domain XLSX/CSV with an advance estimate about two weeks after each month closes. Layer Best Buy API pricing samples on top for SKU-level signal inside a 50,000 calls/day budget.
1,744 datasets. Pick your catch.
What does a demand model actually need that this pool supplies?
A data scientist building an electronics demand model needs three things at minimum: a long target series, an inventory or supply-side companion series to capture the destocking cycles, and enough release-history clarity to backtest without look-ahead bias. The Datadory Computer & Electronics Retail pool — 4 primary datasets plus 10 related records, all free — covers all three.
Which dataset should be your training target?
Granularity runs to six-digit NAICS: 443141 household appliance stores and 443142 electronics stores, monthly since January 1992. Quarterly e-commerce detail extends through Q1 2018 in current detail. Coverage is national, with an experimental Monthly State Retail Sales layer adding modeled year-over-year changes for 50 states plus DC from January 2019.
How do inventories sharpen the forecast?
Seasonal-adjustment variants ship with the release, so you choose adjusted or unadjusted inputs deliberately instead of guessing. The reliability text files document sampling error, which matters when a model treats the latest advance estimate as truth two weeks early and then watches it revise mid-following-month.
Which discovery layers surface extra training corpora?
When NAICS aggregates are too coarse and retailer terms are too tight, three discovery layers in the pool widen the net.
How does the pool rank for forecasting work specifically?
Only the first two belong in the core loop; the rest earn their place as controls and feature extensions.
Pick up where this leaves off
Every one of these ships with sample rows before you commit to anything.
US Census Monthly Retail Trade Survey (MRTS/MARTS) - Electronics & Appliance Stores
Best Buy Developer API - Products, Stores & Categories
U.S. Census Annual Retail Trade Survey / AIES - Motor Vehicle Dealers
NAICS · NAICS_LABEL · GEO_LABEL …+7 more
Google Dataset Search
Hugging Face Datasets Hub - Fashion/Apparel Datasets
downloads · lastModified · siblings
UK ONS Retail Sales Index - Reference Tables Data
Want rows instead of a pitch? Name the datasets.
API, files, or your warehouse. Daily, weekly, or hourly.
Get a sampleQuestions worth asking
Which dataset is used for consumer electronics demand forecasting?
The Census Monthly Retail Trade Survey is the standard baseline: monthly NAICS 443 sales plus matching inventories from January 1992, seasonal-adjustment variants and quarterly e-commerce detail through Q1 2018. Teams add SKU color by sampling Best Buy API pricing and availability, caching responses no longer than 72 hours under the free key.
Is there a long historical time series for electronics retail demand?
Yes. Census MRTS publishes monthly NAICS 443 sales and inventories back to January 1992 as free public-domain XLSX and CSV, split at six-digit depth into 443141 household appliance and 443142 electronics stores. An advance estimate lands about two weeks after month close, with revisions mid-following-month.