Australian cities are transitioning to smart-city hubs. Council masterplans now ask for live foot-traffic, walkable precincts and round-the-clock pedestrian sensing, and the federal Smart Cities Plan requires that the underlying sensors cannot identify a single person. Occivar is being built for that role: a privacy-first occupancy and people-movement layer designed to sit underneath the dashboards a council already runs, not replace them. We are not claiming the backbone yet. We are working towards being it.
Live civic precinct: sample block
A working sketch of what the smart-city data layer looks like at street level: outdoor blocks, a pedestrian road, cars on the carriageway, anonymous dots crossing between the library, town hall, transit hub, markets and civic park. Hover any block, sub-zone or kerb crossing for live counts, dwell and air quality. The side panel mirrors what a council can already publish today: traffic condition, transit status, temperature, footfall, park visitors, air quality, plaza dwell, all from anonymous-on-device sensing.
City of Melbourne already runs a fixed pedestrian-counter network that publishes hourly footfall, proof a council can sense public movement responsibly. The next step every Australian smart-city brief asks for is the same signal at precinct depth: per-zone, live, anonymous, integrated with the systems already in place. These are the three jobs Occivar is designed for.
A single sensor on a streetlight tells you a footfall number. A precinct's worth tells you how the plaza, the library, the transit hub and the park behave together, at minute resolution, across every hour of the day. Anonymous on-device counting means the data layer is honest enough to publish to an open-data portal the same way pedestrian-counter feeds already are.
Every modern liveability index (including the urban-liveability frameworks used to rank Australian cities) weights walkability, access to public open space and active social infrastructure. Until now those were scored on surveys and one-off counts. Continuous, anonymous precinct occupancy turns them into live indicators a council can act on between annual reports.
Density thresholds for civic squares, post-event egress reconciliation, route-choice signals for masterplan reviews, public-space activation studies, all derived from the same anonymous count layer. One sensor stack, many uses, one privacy answer that holds for every one of them.
A people-movement layer earns its place in a smart-city brief by being legible to the systems already in the room: the open-data portal, the BMS, the transport feed, the sustainability report. Four design choices we are committing to early so the architecture is ready when the first deployment is.
Aligned with the federal Smart Cities Plan requirement that streetside sensing cannot identify a person. The model runs on the sensor. The raw frame never leaves it. The platform sees only anonymous counts. Privacy is structural, not a policy promise.
CSV, REST, scheduled emails, Snowflake / BigQuery sync. Designed to feed a council open-data portal at precinct depth, not as a single point, but as a connected layer covering library, plaza, transit and park together.
Reading room, civic square, transit interchange, neighbourhood park: the same edge device, the same privacy answer. One contract, one policy review, one stack that spans the whole precinct rather than a vendor per setting.
Niagara, Honeywell EBI, Slack, Microsoft Teams, ServiceNow, REST webhooks. We do not ask a council to rip-and-replace its stack; we extend the systems running it. Rip-and-replace is how smart-city plans stall; extension is how they ship.
The architecture is the promise.
What the surveillance vendor offers vs what Occivar™ actually puts on a civic plaza, in a library, at a transit interchange or in a neighbourhood park.
Projections, not promises. The numbers below are what published research and existing council pedestrian-counter programs suggest a privacy-first, precinct-scale occupancy layer could deliver, once it's running. We are pre-pilot; these are the targets we are scoping against, not outcomes we have shipped.
Occupancy-led HVAC scheduling consistently shows double-digit energy reductions in published studies; most civic buildings cool empty floors on a calendar, not a count. A precinct-wide live occupancy feed lets the BMS schedule on demand, reported monthly against the council's climate commitments.
Existing council pedestrian counter networks publish a number per sensor per hour. An anonymous precinct layer publishes movement between zones at minute resolution, the same data, two orders of magnitude richer, ready for the open-data portal and the next masterplan review.
Crowd-density safety reviews are usually written after the incident. A live anonymous density signal across plazas and transit interchanges lets the events team and the safety officer act on threshold breaches as they form, and document egress times for the next reconciliation automatically.
Urban liveability frameworks score walkability, open-space activation and social-infrastructure use on annual surveys. Continuous anonymous precinct occupancy turns those sub-domains into live indicators a council can act on between reporting cycles, and benchmark precinct to precinct on the same axis.
Anonymous-on-device means the data is honest enough to publish: to the council open-data portal, to researchers, to the public. The architecture clears the legal review on the first pass because there is nothing identifying to disclose.
The next library brief, the next phone-booth tender, the next bike-rack location, the next pedestrian-priority street, all benefit from a precinct movement layer that has been collecting for twelve months when the brief is written.
Most council sustainability reports separate building energy from building use. A precinct occupancy layer ties them together: energy per visitor, carbon per dwell-minute, scope-2 reductions attributed to demand-led scheduling, all auditable from the same anonymous source.
Sources behind the projection: City of Melbourne pedestrian counting system documentation; published occupancy-led HVAC studies; Australian Urban Liveability Index (Higgs et al.); Australia's Smart Cities Plan anonymization framework. We will publish measured outcomes from each metric once we have a deployment running; until then, every figure on this page is labelled as a projection.
Occivar is backed by the City of Melbourne's Business Growth Innovation Grant (BGIG). We are not running a civic-space deployment yet, and we will not pretend we are. What we are doing is building the architecture a council can adopt without a six-month privacy negotiation, and looking for the first partner to scope a precinct-scale pilot with us.
The role we want is the people-movement layer underneath a council's smart-city stack, feeding the open-data portal, the BMS, the transport feed and the sustainability report from the same anonymous count source. Not a replacement for any of them. A reliable layer they can build on top of for the next ten years.
BGIG grant active. Pre-pilot. First civic deployment targeted H2 2026. Pilot partners welcome: councils, BIDs, precinct authorities and university precincts all in scope.
A civic space cannot deploy a surveillance camera and call it occupancy analytics, and the federal Smart Cities Plan effectively says so out loud. The trade is structural: the model runs on the sensor, the raw frame never leaves it, the platform sees only anonymous counts. There is nothing identifying to retain, disclose or misuse. That is the only architecture that will clear a council legal review on the first pass, which is the only architecture worth basing a smart-city backbone on.
