Operational intelligence
Operational intelligence for emergency response
Firedash reads the incident reports your service already fills in and turns them into three answers: what is happening now, what next week will bring, and what the last eighteen months actually looked like.
Nothing new to collect. No parallel system. No new forms for crews.
Time
16:40
Calls today
7
Open now
0
The problem
Four versions of the same shift
Most coordination problems are not communication problems. They are two people working from different numbers — the radio says one thing, the whiteboard another, the spreadsheet a third, and somebody is holding the rest in their head.
Radio
9
Verbal, unlogged
Whiteboard
11
Current, not shared
Spreadsheet
7
Shared, out of date
Memory
12
Leaves with the shift
All four are correct. None of them agree.
The approach
An intelligence layer over data you already have
Firedash sits between the records a service already keeps and the decisions its officers already make. It adds no data-entry work at either end.
01 · Data
What already exists
Incident reports · public weather warnings · the council’s street-event calendar
02 · Intelligence
What it becomes
Hourly profile · type mix · duration percentiles · the weather effect
03 · Decision
What it supports
“Thursday will need reinforcement” · “this station is carrying the load”
04 · Action
What changes
A roster adjusted two days ahead, not a report filed two months later
Capabilities
See, anticipate, coordinate, learn
A
See
Real-time visibility of the calls still running, which appliances are committed and how much of the service is left. One screen instead of a radio, a whiteboard and somebody’s memory.
B
Anticipate
Expected demand per day from weather warnings and the city’s event calendar. The arithmetic is on screen — baseline, weather, weekday, events — because a forecast nobody can check is a forecast nobody acts on.
C
Coordinate
Command and field teams reading the same incident log, the same map and the same resource count, at the same moment.
D
Learn
Eighteen months of activity as a working instrument: type mix, seasonality, time on scene, and the hour-by-hour pattern that shows when a shift runs short.
The product
This is the interface, not an illustration
Every screen below is the running product, rendered from the same dataset the demo uses. No mockups, no stock dashboards.
Control room
A shift replayed minute by minute
Forecast
Seven days, with the model in full view
- Tue 1711
- Wed 1812
- Thu 198
- Fri 2011
- Sat 219
- Sun 2213
- Mon 238
Activity
Eighteen months of pattern
Simulated data
Evidence
Measured, not asserted
Firedash was calibrated on 4,519 real interventions recorded by the Bilbao Fire Service between 2019-01-01 and 2020-05-31. These are findings from that work.
+38%
when two warnings overlap
Days with two or more simultaneous weather warnings averaged 11.59 calls against 8.42 on a clear day. Measured over 96 days.
No change
from a single warning
One active warning averaged 7.74 calls — slightly below baseline. A forecast that only reported good news would not be worth acting on.
2 in 3
calls are technical assistance
Wasp nests, façade inspections, pumping, door openings. Fires are why the service exists and a minority of its work — and rosters are planned against the whole mix.
45 min
median time on scene
Half of all calls are over inside three quarters of an hour. The long tail is what ties up an appliance for a whole afternoon.
An association measured over 96 days, not a causal model. Enough to move a duty roster, not enough to publish in a journal — and the difference matters.
Interactive demo
Don’t read about Firedash. Open it.
A full shift you can scrub through minute by minute, a seven-day forecast with its arithmetic exposed, and eighteen months of activity. Runs in the browser, takes about two minutes, nothing to install.
Launch the interactive demoTry it here
Time
16:40
Calls today
7
Open now
0
Drag the timeline. The map moves with it. 0 open
Why Firedash
Built for emergency operations, not adapted to them
Data and AI, applied narrowly
Not a general analytics tool pointed at emergencies. Built around the way a fire service actually records its work.
Operational, not reporting
The output is a decision a duty officer takes this week, not a dashboard reviewed next quarter.
Fits what you already run
Assets, hazards and measures follow the NGSI-LD smart data models used across European emergency projects, so the output can travel to regional systems.
No new data collection
Firedash reads the fields already filled in after each call. Adoption does not depend on crews changing how they work.
Developed by Galde
A data and AI team working with reinsurers, public administrations and platform vendors across Europe.
Grant agreement
951981
European validation
Selected and funded through Horizon 2020
Firedash was built inside the REACH Incubator, the European data incubator funded by the European Union’s Horizon 2020 research and innovation programme. The product was developed with the Bilbao Fire Service on its own operational records.
The project at REACH IncubatorThis project has indirectly received funding from the European Union’s Horizon 2020 research and innovation programme under project REACH Incubator (Grant Agreement no. 951981).
See what Firedash could look like in your operation
There is nothing to buy off a price list. If you are preparing a tender, or trying to work out what your own records could tell you, that is the conversation worth having.