Exposure records preserved by two users of Japan's COVID-19 Contact-Confirming Application (COCOA) during the seventh wave: data and reproduction script

Companion dataset for: Nakagawa S, Kumagai S, Yamamoto A. Exposure Records Preserved by Two Users of Japan's COVID-19 Contact-Confirming Application During the Seventh Wave: A Descriptive Study (medRxiv, doi:10.64898/2026.05.06.26352506, version 2; submitted to a peer-reviewed journal). Source of the exposure and key records: the authors' own devices (COCOA 2.0.1 "Save information" export and the iOS Exposure Checks log). Reported case counts are public data of the Tokyo Metropolitan Government and Kyoto Prefecture. No third-party personal data are included. Files: exposure_windows_observer1.csv (367 rows): one row per GAEN ExposureWindow in Observer 1's export, exposure dates 2022-03-29 to 2022-11-08. exposure_windows_observer2.csv (261 rows): same for Observer 2, 2022-04-03 to 2022-09-24. keys_checked_daily.csv (478 rows): Diagnosis Keys checked and matched per day, from the iOS Exposure Checks log of each device. prefecture_daily_cases.csv (276 rows): daily new cases, Tokyo and Kyoto Prefecture, 2022-03-01 to 2022-11-30. daily_series.csv (407 rows): derived daily series used for Figure 1. reproduce.py: recomputes the numbers in Tables 2–3, the correlations, the attenuation distribution and the before/after comparison from the raw files (Python 3, pandas, numpy). README.md: column definitions and notes. CITATION.cff: citation metadata. One window represents up to 30 minutes of reception for one Diagnosis Key; windows are neither people nor encounters. Analysis periods: Observer 1, 2022-03-29 to 2022-10-22; Observer 2, 2022-04-03 to 2022-10-18.