Checking access…
Checking access…
High-resolution location intelligence derived from network mobility signals — delivering crowd movement analytics, origin-destination flows, venue attribution, and dwell-time scoring for urban planning, retail intelligence, and public safety applications. Built on a scalable GCP spatial pipeline with 100m – 300m precision.
Omni Pinpoint transforms raw network mobility signals into a full-stack geospatial intelligence platform. Eight analytical modules — from population density heatmaps and origin-destination flow modelling through to AI-driven spatial querying — share a single BigQuery spatial pipeline and deliver insights at sub-300m H3 resolution across any geography.
Omni Pinpoint is powered by three datasets: a high-frequency Footfall feed capturing device-level location events, a Master Dataset resolving raw coordinates into a full geographic hierarchy with spatial metadata, and a Demographics feed enabling crowd composition analysis by age, gender, and segment.
Footfall events are spatially joined against the Master Dataset in BigQuery to resolve raw coordinates into place-aware records, then enriched with demographic attributes before writing to the spatial output tables delivered via Analytics Hub.
Omni Pinpoint processes 10 TB/day of raw location pings from 70M subscribers (~500 rows/sub/day, ~35B rows/day total), spatial-joins with the 200 MB Master Dataset and 4M-row Demographics table, then aggregates to 1.2M POI-level hourly and daily summary tables. Hourly tables use a configurable rolling window; daily tables retain 12 months. The dominant cost is daily batch compute on BigQuery Flex Slots — storage is secondary.
BQ Flex Slots compute (gold) stacked with storage cost (teal) · 18% YoY growth · compute:storage ~195:1 at baseline
BigQuery Flex Slots is the recommended platform: a single daily SQL + GEO batch job with no persistent compute, no Spark overhead, and native H3 spatial support. On-demand BQ pricing at $6.25/TB would cost ~$62/day just for the raw 10 TB scan alone — Flex Slots burst capacity eliminates per-query billing entirely during the daily window.
Burst capacity activated only during the 6-hour daily processing window. At 35B rows/day, H3 coordinate conversion is a compute-heavy UDF pass on every ping, and the 70M-subscriber demographic join is a large shuffle — together requiring a minimum of 500 slots for nightly completion. sub_hash is the pseudonymised join key across all three input datasets. Native BQ GEO functions and H3 extensions handle all spatial operations — no external Spark cluster needed. The 200 MB Master Dataset is broadcast-joined, making the spatial lookup step highly efficient.
Slot count × daily window × $0.04/slot/hr determines the full compute cost. At 35B rows/day, the H3 coordinate-to-hexagon conversion (a compute-intensive UDF operation on every ping) combined with the 70M-subscriber demographic shuffle-join makes 500 slots the practical minimum for a 6-hour nightly window. At 200 slots the same pipeline requires 12–15 hours — incompatible with a daily cadence. Cost scales linearly — halving slots roughly doubles the required window, keeping daily cost identical.
Each day produces ~9.2 GB of hourly aggregation output (1.2M POIs × 24 hrs × 16 demo dims) and ~384 MB of daily rollup. Both tables use rolling partition expiry — old days auto-delete as new ones arrive. BQ's blended active/long-term pricing brings total aggregated storage to ~$224/year at the default ~969 GB steady-state — under 0.6% of the annual compute cost at 500 slots.
| Table | Rows/day | GB/day | Active (≤6mo) | Dormant (>6mo) | $/yr |
|---|---|---|---|---|---|
| Hourly (POI×hr×demo) | 461M | 9.2 GB | 829 GB | — | $199 |
| Daily (POI×day×demo) | 19M | 0.4 GB | 69 GB | 71 GB | $25 |
| Total | — | 9.6 GB | 898 GB | 71 GB | $224 |
Annual compute scales with subscriber volume growth at the configured rate. Storage grows proportionally with data volume. Storage cost remains under 1% of the total throughout the forecast period — this is a compute-dominated model.
| Year | Daily Compute | Annual Compute | Hourly SS (GB) | Daily SS (GB) | Storage Cost / yr | Total |
|---|---|---|---|---|---|---|
| 5-Yr Total · Yr 5 SS↑ | — | — | — | — | — | — |
Plain-English guide to all cost models, assumptions, strategies, and caveats across every product page.
