Slide 2
@bposton
Ben Poston analyzed police response
times in Milwaukee
Slide 3
@palewire
Ben Welsh analyzed fire and
medical response times in L.A.
http://www.latimes.com/lafddata
Slide 5
WHAT FIREFIGHTERS REALLY DO
● More than 80% of LAFD Fire vs. medical responses at LAFD
responses are medical
● Media coverage and fire
department PR tend to
focus on fires
● Extremely long calls get
good anecdotal
coverage, but full system
rarely analyzed (This pie contains real data!)
Slide 7
A layman's definition
How long it takes from when you
dial 911 until a rescuer arrives at the
scene.
Slide 8
HOW HARD
CAN THIS
BE?
STEVE LOPEZ
@latimes columnist Source: http://lat.ms/11DUUNB
Slide 10
WHY
● Your call is passed between different departments.
● Each can have its own database, analytics, politics.
● Often, nobody responsible for tracking the whole thing.
● People who could have few incentives, little training.
Slide 11
A SIMPLIFIED VERSION
HOLD CALL
TIME PROCESSING
TIME
TURNOUT TRAVEL
TIME TIME
Slide 12
HOLD TIME
● How long it takes for
a dispatcher to pick
up the 911 call.
Slide 13
CALL-PROCESSING TIME
● How long it takes the
dispatcher to log the
incident and send
help.
Slide 14
TURNOUT TIME
● How long it takes
rescuers to suit up
and get ready.
Slide 15
TRAVEL TIME
● How long it takes
rescuers to drive to
the scene.
Slide 16
A REAL-LIFE VERSION
POLICE CALL
HOLD TIME
PROCESSING
POLICE POLICE
TURNOUT TRAVEL
FIRE CALL FIRE FIRE
PROCESSING TURNOUT TRAVEL
Slide 17
AT THE CORE, A PUBLIC DATABASE
Slide 18
INCIDENT TABLE
● One record per call
● Unique identifier
● Incident type
● Dispatch address
● Lots of timestamps
Slide 19
RESPONSE TABLE
● One record per unit
● Many to one with call
● Describes unit type
● More timestamps
Slide 20
WHAT'S
YOUR
STANDARD?
BRANT HOUSTON
Journalism professor Source: My memory
Slide 21
CLEAR STANDARDS
BUT THEY ARE OPTIONAL
Slide 22
HOLD TIME
● How long it takes for
a dispatcher to pick
up the 911 call.
● NENA 56-005 says
90% of calls should
be answered in
under 10 seconds
Slide 23
CALL-PROCESSING TIME
● How long it takes the
dispatcher to log the
incident and send
help.
● NFPA 1221 says 90%
of fire and medical
calls should be
dispatched in under
60 seconds
Slide 24
TURNOUT TIME
● How long it takes
rescuers to suit up
and get ready.
● NFPA 1710 says 90%
of medical units
should be moving in
under 60 seconds.
Slide 25
TRAVEL TIME
● How long it takes
rescuers to drive to
the scene.
● NFPA 1710 says the
first fire or medical
unit should arrive in
under 4 minutes for
90% of responses.
Slide 26
Do local firefighters
meet the standards?
Are times getting
faster or slower?
Does everybody get
the same service?
You
Professional observer
Slide 34
Time isn't everything
Most calls are not
emergencies
Grandma gets
lonely
GENERIC FIRE CHIEF
Anytown, USA
Slide 35
FOCUS ON CARDIAC ARRESTS
● Only ~2% of medical
calls, but everyone
agrees time imperative
● Brain death can begin
after as little as four
minutes
● Recent push for
increased speed from
scientists and doctors
Slide 37
YOU NEED A TEAM
@LAJourno @katelinthicum
Investigative reporter, City Hall beat writer,
nightblogger fears no deadline
Slide 38
AND A FEW RABBIS
INSIDER DATA GEEKS REAL FIREFIGHTERS
Every database has quirks, They live where your data
they know them meets reality
Slide 39
THE END
All materials at palewi.re/talks
Slide 40
"RESPONSE TIMES
DON'T MATTER"
(And other
excuses police will
make)
By Ben Poston
Slide 41
- Let's say you get police response time
data
- Your PD will likely say response times
don’t matter.
- They will say that they are not a good
measure on which to judge a PD’s
performance.
Slide 42
How come?
Research done in the 1970s and 1980s that
showed that a rapid police response to
calls for service only makes a difference
on crimes in progress and victims often
don't report crimes right away.
The research suggested that just rushing
to calls depletes valuable resources that
could be used for crime prevention.
Slide 43
Milwaukee PD: A case study
- MPD crime numbers
had declined for 4
straight years
- We heard skepticism
from every corner of
the community
- I wanted to get behind
the numbers
Slide 44
Response times lag as strategy shifts
I reported response times were
10% slower on average and
they had gotten slower in 13
of 15 major categories.
Examples: Officers took more
than 2 hours to respond to a
fatal car crash, 45 minutes for
a fatal stabbing and more
than a half-hour for an armed
robbery
Residents complained that
cops don’t respond at all to
minor crimes like theft from
auto, drug dealing, graffiti,
and loud music.
Slide 45
Police response: "Meh"
The department’s take:
Response times don’t matter
that much.
"Since the average resident of
this city is willing to wait four
hours for the cable guy and
half a day for a furniture
delivery, it seems to me a
reasonable delay in responding
for a call is an acceptable
balance," the police chief said.
Slide 46
That's not all that changed ...
The department’s entire
strategy had shifted from a
reactive strategy to more of a
proactive one.
Less focus on rapid response
and instead quadrupled the
number of cars stops and
tripled the number of subject
stops. They were all about
“high visibility policing.”
The police said this
deployment shift was a key
driver to their declining crime
numbers
Slide 47
Calls are prioritized
Like most big-city PDs, Milwaukee police now prioritize their calls, using
a system from Priority 1 to 5. For minor calls like loud music and drug
dealing, police often take reports over the phone or send an officer the
next day to prioritize.
They called it the differential police response unit.
In Milwaukee, the number of these had more than doubled in 3 years,
raising lots of questions.
Slide 48
Tips for reporting on police
response times
Hit the meetings: go to
neighborhood watch block
meetings.
This is where people complain
every week about crime and
poor police responses.
Residents also complained
about feeling harassed by
police dispatchers - they
claimed they bullied them or
talked them out of reporting a
crime.
Slide 49
Spend a night in a hood:
I had been talking to resident
Thom Reed, who lived in a
high-crime neighborhood.
The things he told me were
hard to believe, so I went out
to see it myself.
In one six-hour period, I
witnessed shots fired and
drug dealing. It helped me
write with authority and add
color to a follow-up piece.
Slide 50
Differential police response units
Giving a lower priority to less
serious calls makes sense, but this
policy upset residents more than
anything.
Most people have expectation of
seeing a police cruiser when they
call 911. After a few hours, they get
frustrated and give up and this
colors their view of police.
Skeptics claimed that police used
the DPRs to slough off crime,
though we never could prove that.
Slide 51
KNOW YOUR DATA
I initially got police
response time data and
realized later that it was
from time entered to
dispatched, but didn’t
capture the time to arrive
on scene, making the
times seem faster.
Slide 52
GET THE RAW DATA
We got aggregated data for the 2011 stories, but later
got the raw CAD data that allowed us to dig deeper for
trends. You can also use that call data to populate an
interactive maps of calls.
Slide 53
Track crime
through the
criminal justice
system
CAD data is the very beginning of the process.
With police CAD data, you can request the associated
incident report number. This would allow you to see
how often police create a report by call type and
overall.
You can also see how often police flag calls as
nuisance calls.
Slide 54
LINK 'EM UP!
If you were so inclined, you could then link those
incident reports up with crime data if your police
department is a NIBRS agency or collects
incident-level data.
From there, you merge that data with your District
Attorney’s case management data (assuming your
DA is not living in the stone ages, which many are).