Datasets:

EPI-Eval
/

global-mobility

pretty_name: Google Community Mobility Reports — global daily
license: cc-by-4.0
size_categories:
  - 10M<n<100M
tags:
  - cadence-daily
  - geo-global
  - surveillance-mobility
  - tier-2
  - availability-inactive
schema_version: '0.1'
source_id: global-mobility
source_url: https://www.google.com/covid19/mobility/
manifest_section: §15.7
surveillance_category: mobility
pathogens: []
availability: inactive
availability_notes: >-
  Google ended data collection 2022-10-15. Historical CSV remains available; not
  updated.
access_type: csv
tier: 2
cadence: daily
geography_levels:
  - national
  - subnational-state
  - subnational-county
  - subnational-city
geography_countries:
  - multiple
gold_standard_for: []
vintaged_version_of: null
succeeds: null
derived_from: []
value_columns:
  - name: retail_and_recreation
    unit: percent change vs. baseline
    value_type: index
    description: |
      Daily % change in visits + length of stay at retail / recreation places
      (restaurants, cafes, shopping centers, theme parks, museums, libraries,
      cinemas) compared to the baseline (median value for the 5-week period
      Jan 3 – Feb 6, 2020).
    aggregation: mean
  - name: grocery_and_pharmacy
    unit: percent change vs. baseline
    value_type: index
    description: >-
      Daily % change in visits to grocery markets, food warehouses, farmers'
      markets, specialty food shops, drug stores, pharmacies.
    aggregation: mean
  - name: parks
    unit: percent change vs. baseline
    value_type: index
    description: >-
      Daily % change in visits to local parks, national parks, public beaches,
      marinas, dog parks, plazas, public gardens.
    aggregation: mean
  - name: transit_stations
    unit: percent change vs. baseline
    value_type: index
    description: >-
      Daily % change in visits to public transport hubs (subway, bus, train
      stations).
    aggregation: mean
  - name: workplaces
    unit: percent change vs. baseline
    value_type: index
    description: Daily % change in visits to places of work.
    aggregation: mean
  - name: residential
    unit: percent change vs. baseline
    value_type: index
    description: |
      Daily % change in time spent in places of residence. (Note: this is
      length-of-stay, not visit count — different metric than the other five.)
    aggregation: mean
notes:
  extra_columns:
    - column: location_name
      description: |
        Most-specific populated source field — county name for county rows,
        metro name for metro rows, ISO sub-region name for state rows, country
        name for national rows. `location_id` is canonical.
  interpretation_caveats:
    - column: residential
      caveat: |
        Residential is reported as % change in *length of stay*, not visit
        count. The other five categories are visit-based. Don't sum or compare
        directly across these two flavours.
    - column: parks
      caveat: |
        Parks shows extreme seasonal cyclicity (winter vs. summer is a 200%+
        swing in many countries) that reflects normal seasonality, not
        pandemic-period behaviour change. Detrend before any pandemic analysis.
    - column: workplaces
      caveat: |
        Holidays / weekends are baked into the baseline since the baseline is
        a median across weekdays + weekends. Day-of-week cycles in the output
        are real but reflect *deviation* from the comparable weekday's baseline.
  general: |
    Aggregated, anonymised mobility signals derived from Google Maps Location
    History data. Coverage is global at the national level; subnational depth
    varies by country (US has county granularity, most others have ISO 3166-2
    state-level only). Privacy threshold means small geographies / low-traffic
    days are NaN.
extra:
  baseline_period: median value across Jan 3 – Feb 6, 2020 (5 weeks, per weekday)
  collection_end_date: '2022-10-15'
  source_columns_omitted: |
    `place_id` (Google internal place identifier — opaque), `sub_region_1`,
    `sub_region_2`, `country_region`, `metro_area` (collapsed into the derived
    `location_id` + `location_name`), `country_region_code`, `iso_3166_2_code`,
    `census_fips_code` (likewise collapsed).
computed:
  last_ingested: '2026-04-26T12:14:44Z'
  row_count: 11728497
  time_coverage:
    - start: '2020-02-15'
      end: '2022-10-15'
  geography_unit_count: 5306
  observed_cadence_days: 1
  missing_gaps: []
  data_hash: 63d11fb60ba0adbb

Google Community Mobility Reports — global daily

Aggregated, anonymised mobility signals derived from Google Maps Location History data. Coverage is global at the national level; subnational depth varies by country (US has county granularity, most others have ISO 3166-2 state-level only). Privacy threshold means small geographies / low-traffic days are NaN.

Source: https://www.google.com/covid19/mobility/

Coverage

Time: 2020-02-15 → 2022-10-15
Cadence: daily (observed median spacing: 1 days)
Geography levels: national, subnational-state, subnational-county, subnational-city — 5306 unique location IDs
Countries: multiple
Pathogens: —
Surveillance category: mobility
Rows: 11,728,497

Columns

Column	Unit	value_type	Aggregation	Description
`retail_and_recreation`	percent change vs. baseline	`index`	`mean`	Daily % change in visits + length of stay at retail / recreation places
(restaurants, cafes, shopping centers, theme parks, museums, libraries,
cinemas) compared to the baseline (median value for the 5-week period
Jan 3 – Feb 6, 2020).

`grocery_and_pharmacy`	percent change vs. baseline	`index`	`mean`	Daily % change in visits to grocery markets, food warehouses, farmers' markets, specialty food shops, drug stores, pharmacies.
`parks`	percent change vs. baseline	`index`	`mean`	Daily % change in visits to local parks, national parks, public beaches, marinas, dog parks, plazas, public gardens.
`transit_stations`	percent change vs. baseline	`index`	`mean`	Daily % change in visits to public transport hubs (subway, bus, train stations).
`workplaces`	percent change vs. baseline	`index`	`mean`	Daily % change in visits to places of work.
`residential`	percent change vs. baseline	`index`	`mean`	Daily % change in time spent in places of residence. (Note: this is
length-of-stay, not visit count — different metric than the other five.)

Additional data columns

location_name — Most-specific populated source field — county name for county rows, metro name for metro rows, ISO sub-region name for state rows, country name for national rows. location_id is canonical.

Interpretation caveats

Things that may differ from how other sources define a similar measure. If you're combining this dataset with another, read these first.

residential — Residential is reported as % change in length of stay, not visit count. The other five categories are visit-based. Don't sum or compare directly across these two flavours.
parks — Parks shows extreme seasonal cyclicity (winter vs. summer is a 200%+ swing in many countries) that reflects normal seasonality, not pandemic-period behaviour change. Detrend before any pandemic analysis.
workplaces — Holidays / weekends are baked into the baseline since the baseline is a median across weekdays + weekends. Day-of-week cycles in the output are real but reflect deviation from the comparable weekday's baseline.

Access

Availability: inactive
Access type: csv
License: cc-by-4.0
Tier: 2

Schema version 0.1 · Last ingested 2026-04-26T12:14:44Z · source_id: global-mobility · Manifest section §15.7