When someone calls 911, the call taker logs it in the agency's computer-aided dispatch (CAD) system: what was reported, where, when, and which units were assigned. That CAD record is the raw material for everything you see on Dispatchr.
Agencies choose to publish
Many agencies publish a public view of their CAD activity through open-data portals — among them the Toronto Police Service and Toronto Fire Services, whose feeds power the Toronto pages here. Publication is a transparency choice: it lets residents see what their emergency services are responding to, and it lets researchers and journalists work with real operational data instead of anecdotes.
What gets filtered before publication
The public feed is not the full CAD record. Agencies pre-filter for privacy before anything is published: certain sensitive call categories are withheld entirely, exact addresses are generalized (typically to an intersection or block), and personal details are never included. What remains is a call type, a time, an approximate location, and limited operational fields. Dispatchr applies further aggregation on top of that for these Explore pages — counts and patterns only, no individual events.
Publication schedules vary
Each agency publishes on its own schedule — some update their feeds frequently through the day, others in batches. That is why Dispatchr avoids promising any specific data freshness, and why Explore pages work with completed calendar months: by the time a month is aggregated, every agency's feed for that period is settled.
What Dispatchr adds
Raw agency feeds differ in format, naming, and geography. Dispatchr ingests them, harmonizes the call types, assigns priority levels, resolves locations against official neighbourhood boundaries, and flags records with data-quality problems so they don't distort statistics. The result is one consistent dataset across cities that were never designed to be compared.