Trillboards
Menu
Where they went
Geohash activity, hourly
The camera readings of place activity, placed where the vehicle was when each reading closed: on the map-matched route where one covers that moment, otherwise at the screen's last GPS fix, with the age of the fix. One row per geohash-6 cell, about 1.2 by 0.6 km, per hour.
The question it answers
What is happening on this block, this hour?
10,098
ZIP areas on the camera fleet's GPS trail
154
DMAs on the trail
88%
of the US population lives in those DMAs
Measured over the 90 days to September 25, 2026.
The dataset
- Grain
- Hourly.
- Cadence
- Hourly rows, one file per day.
- History
- From June 24, 2026.
- Format
- Parquet or JSONL.
- Delivered to
- Your S3 bucket, your GCS bucket, or a time-limited download link, with a manifest and a schema.
- Method
- How each measure is produced and how to add rows up: the measurement methodology.
What buyers do with it
- Street-level activity indexes for retail, real estate and local economies.
- Demand around venues and events, hour by hour.
- A geographic join to your own data, with no screen ids involved.
The table
Read from the export registry each time this site is built.
geohash_activity_hourly
Hourly, from June 24, 2026- Grain
- One row per hour_start and geohash6.
- Delivery
- One file per day, as Parquet or JSONL.
- Basis
- Measured.
- Definition
- Version 1.
| Column | Type | What it holds |
|---|---|---|
hour_start | timestamp | Start of the UTC hour. A camera window counts in the hour it closed. |
geohash6 | string | Geohash-6 cell (about 1.2 x 0.6 km) where the screen was when the window closed. |
observation_windows | integer | Camera-graded windows placed in the cell. Unit: windows. |
face_backed_windows | integer | Of those, the face-backed windows. Unit: windows. |
person_backed_windows | integer | Of those, the person-backed windows. Unit: windows. |
screens | integer | Distinct screens whose windows the cell holds. Unit: screens. |
avg_face_count | double | Mean faces per window: the device's face_count where it wrote one, else the window's mean face count. Unit: faces_per_window. |
p95_face_count | double | Exact 95th percentile of the per-window face reading over the cell-hour. Unit: faces_per_window. |
path_positioned_windows | integer | Windows placed on the map-matched vehicle path at their instant. Unit: windows. |
last_fix_positioned_windows | integer | Windows placed at the screen's last GPS fix at or before their instant. Unit: windows. |
fix_age_p50_s | integer | Median age of those fixes at the window's instant. Unit: seconds. Null: No window of the cell-hour was placed by a fix. |
fix_age_p90_s | integer | 90th-percentile age of those fixes. Unit: seconds. Null: No window of the cell-hour was placed by a fix. |
Request a sample
Tell us the markets and the weeks you want. The sample is cut from production by the same export that makes the full delivery.
Request a sample