Frozen 14 October 2026. The competition-radius test replaced 123 of 332 shortlisted tract areas (37.0%) across 3,277 residential anchors in eight selected counties. It is model sensitivity, not predicted sales, profitable-site rankings or a national sample.
# Reproduce the frozen US coffee sensitivity study Study: https://www.locusintel.io/blog/us-coffee-expansion-shortlist-robustness Source cutoff: Overture 2026-09-23.1; ACS 2020–2024 five-year; 2024 Gazetteer. This is an eight-county model-sensitivity experiment, not a current-venue census or a profitable-site recommendation. 1. Extract `calculation-package.zip` into a new directory, preserving its paths. 2. Put the separately downloaded `frozen-inputs.zip` inside `data/us-retail-expansion-study/2026-10-14/` in that directory. 3. With Python 3.12 and NumPy, pandas and Pillow installed, run from that directory: ``` python scripts/us-retail-expansion-study/restore_frozen_inputs.py python scripts/us-retail-expansion-study/validate.py python scripts/us-retail-expansion-study/test_analysis.py ``` `analyze.py` and `audit_business_categories.py` can re-run the saved-data calculations. No network, API credentials or paid provider is needed. The restoration helper verifies all 44 original input hashes and refuses to overwrite differing files. Current results include the exact issuer-confirmed closure exclusion and locked-category repair. The two named pre-correction directories are historical, not the current baseline. The saved protocol distinguishes pre-result decisions from post-result diagnostics. Source notices and licences remain attached. The evidence manifest records public download hashes; frozen-inputs-manifest.json records each original input's hash.
# Frozen study source notices Overture Maps Foundation, overturemaps.org. Data accessed 14 October 2026; Places release 2026-09-23.1. Attribution and source terms: https://docs.overturemaps.org/attribution/ and https://docs.overturemaps.org/guides/places/. Per-record provider, source identifier, update time and license fields are preserved in the unchanged raw extract. Providers include Meta, Microsoft, PinMeTo, Krick, RenderSEO, DAC and BrightQuery under CDLA Permissive 2.0; Foursquare under Apache 2.0; AllThePlaces under CC0 1.0, as applicable to each source entry. The accompanying `licenses/` contains full license text and a complete current Foursquare notice snapshot. Overture's source attribution specifies: Copyright 2024 Foursquare Labs, Inc. All rights reserved. Foursquare data was transformed to the Overture schema; changed 2026-03-18. Preserve that attribution along with the current issuer NOTICE. Locus transformations: selected eight saved extraction boxes and the food_and_drink hierarchy; selected the recorded source fields; combined records by exact ID without conflicting duplicates; derived tract-level counts, ranks and illustrative shortlist sensitivity. Original source rows and their provenance were not edited. Issuer-confirmed closure exclusions and the restored locked taxonomy rule apply to calculations through separate files, not alterations of source data. Computed pre-correction histories remain explicitly labelled historical. No claim of venue completeness, independent current-status verification of every record, sales prediction or endorsement by a source provider. U.S. Census Bureau sources: public 2020–2024 ACS five-year table-based Summary Files B01003/B19013 and the aligned 2024 Gazetteer. The originally collected 2025 Gazetteer is superseded and retained only for audit history. Exact source URLs/hashes and exclusions are in the manifests. Census inputs were not substituted with 2026 observations. Source site: https://www.census.gov/programs-surveys/acs/data/summary-file.2024.html. The compressed frozen-input bundle preserves original bytes and hashes. Compression is storage packaging, not a new data capture or a change in the study population.
