Business Analytics Overview

advertisement
Improve public safety and security
© 2012 IBM Corporation
UK Ministry of Justice
Predicting offender behavior and reducing recidivism
The need:
The U.K. Ministry of Justice needed a way to analyze vast amounts
of crime and offender data to understand which proactive measures
would be likely to prevent recidivism.
The solution:
The ministry turned to IBM SPSS predictive analytics software which
analyzes both numeric and textural data – to analyze millions of
prisoner files. The analysis is helping them develop treatment targets
for prisoners throughout their sentence to reduce the probability they
will commit crimes upon their release. The Ministry of Justice now
has the ability to perform a complex analysis of millions of files to
reveal trends and patterns hidden within the data, has developed
more effective treatment targets for prisoners throughout their
sentence, and has a greater understanding of which interventions are
likely to prevent recidivism .
"IBM predictive analytics technology
provides us with valuable insight into
offender data which helps us to
predict who may reoffend and
enables us to advise on preventative
measures before a prisoner’s
release date."
- Spokesperson for the Ministry of
Justice
Solution components:
 IBM SPSS
What makes it smarter:
 Contributes to the ultimate objective of improving public safety.
 Violent crime recidivism prediction improved from 68 to 74 %
 General offenses recidivism prediction improved from 76 to 80%
© 2012 IBM Corporation
City of Lancaster
The need:
Between 2000 and 2010, the population of this U.S. city grew by more than
30 percent – and the crime rate rose along with it. The City of Lancaster
wanted to understand the root causes of criminal activity, gain visibility into
when it might occur and learn how to fight it more effectively.
The solution:
The last thing criminals want to be is predictable. Yet this city has found that
crime can be as foreseeable as an afternoon thunderstorm. It is using
predictive analytics to forecast when and where crimes are likely to occur, so
the sheriff's department can focus law enforcement resources on expected
hot spots. Complex statistical models are applied to a database of crime
reports, arrest records and 911 calls, producing a color-coded, geographic
map that anticipates criminal activity across the metropolitan area. This
monthly forecast acts as a guide for crime-fighting strategies. In addition, the
city gains surprising and valuable insights into the nature of crimes. For
example, analysis revealed that a large proportion of property-related crimes
could be attributed to careless owners - people who forget to lock their cars or
close their garages. So the city launched a public education campaign,
encouraging people to pay attention and avoid the crimes altogether
“… We use a combination of current
and historical crime data to predict
the timing and location of crimes,
much as a meteorologist uses
weather data to forecast
temperatures and rain patterns.”
– Senior Criminal Justice Analyst
Solution components:
 IBM SPSS
What makes it smarter:



Delivers predictive insight into probable crime event patterns, helping to
deploy sheriff’s department resources more effectively across the city
Uses intuitive graphics to help everyone from senior officials to police
officers on the beat understand crime analyses at a glance
Enables detailed analysis of current and historic crime data to understand
root causes and support the creation of crime prevention campaigns
© 2012 IBM Corporation
Greater Manchester Police
The need
Greater Manchester Police (GMP) is the largest police force in the United
Kingdom, monitoring 500 square miles and a population of 2.5 million people.
Consultation is a statutory function of the Greater Manchester Police (GMP)
and is ongoing throughout the year. As part of the National Reassurance
Project GMP sent a survey to 9,000 households in order to establish a
community-based policing program. The concept focused on identification of
“signal crime.” GMP had been using a low-level paper system and Microsoft
Excel software to track the program. This method proved unfruitful because of
Excel’s limited analysis capabilities. GMP sought a solution that would help it
survey the 9,000 households more accurately and efficiently.
“Not only did we benefit from a
massive reduction in person time
for this project…but these results
are now being referenced by other
Greater Manchester Police
divisions as contributions to
‘efficiency savings’ required by
government.”
-- Keith Bentley, chief
superintendent of operations,
Oldham division, GMP
The solution
GMP decided to adopt the IBM SPSS software analytical solution. By using
this sophisticated software, GMP aimed to gather a wide range of public views
on area policing that could be incorporated into the National Reassurance
Project. The questionnaires were seamlessly created within the system, and
the responses were scanned electronically, eliminating the need for manual
data entry.
Solution components:
 IBM SPSS
What makes it smarter



Saved approx. EUR20,000 within the first two weeks
Industry compliance was also achieved through “efficiency savings”
required by the government.
Increased survey and analysis capability impressed several partners who
have now fully bought into the National Reassurance Project in Oldham
© 2012 IBM Corporation
Groningen Regional Police Force
The need
Groningen Regional Police Force is required to achieve a variety of government
targets. One of the targets is related to citizens’ satisfaction with the service they
receive from the police. The Groningen Regional Police Force needed to improve
citizen satisfaction levels to meet the Dutch government’s new targets. However, its
existing paper-based satisfaction surveys were expensive to run, could not be
tailored to different groups of respondents, and often failed to provide a
representative sample of the local population.
The solution
The Groningen Regional Police Force built a sophisticated online satisfaction
survey using IBM SPSS that allows the department to make the surveys as simple
or complex as it chooses. IBM automatically tailors questionnaires to a respondent’s
situation. Whether he or she is a citizen, a victim, an offender or a suspect, the
survey can adjust accordingly, eliminating unnecessary questions and getting the
most out of each participant’s answers. Finally, they implemented IBM SPSS Data
Collection Web Interviews software to be made available on multiple channels,
reaching as many people at one time as possible.
were derived.
“Our citizen satisfaction survey
project won an innovation award
from the MOA, the Dutch Centre
for Marketing Intelligence and
Research. But the best testament
to our success is the fact that both
the citizen and employee surveys
are being adopted by other police
forces around the country. "
- Johan Huizing, Project
Manager Quality Research,
Groningen Regional Police
Force.
Solution components:
 IBM SPSS
What makes it smarter



Eliminating paper questionnaires, the department reduced printing,
postage, scanning and data validation costs
Better citizen satisfaction
Early warning triggers of PTSD
© 2012 IBM Corporation
Hessian Police Criminalistic-Criminological Research Center
Investigating serious offenders with IBM SPSS
The need
Hessian Police Criminalistic-Criminological Research Center (KKFoSt) was
investigating the criminal biographies of repeat and serious offenders, referred as
MITs. In 2007, 1,502 MITs were registered in Hessen, having committed a total of
5,673 criminal offenses. The aim of the KKFoSt project is to use experience to
optimize the way in which MITs are dealt with and prosecuted - for example, by
designing tailor-made preventive measures, combining resources, or investigating
those who have a tendency toward criminal activity. They needed a solution that
could provide a way to identify every criminal offender and organize them by the
kinds and numbers of crime they committed.
“For us, it is the standard solution
for professional statistical analyses.
We have had confidence in the
solution for many years and have
implemented it in practice. "
- Dr. Claudia Koch-Arzberger,
Head of the KKFoSt.
The solution
To complement the quantitative database of criminal records, the KKFoSt
conducted qualitative interviews with selected MITs. They analyzed the data
collected and translated it into meaningful knowledge for the police, the using IBM
SPSS software. This enabled researchers to preprocess the structured data from
the full MIT survey and the unstructured information from the interviews so it can be
processed using univariate, bivariate and multivariate analytical methods. Key
findings for the police were extracted from the connections and pattern. Cluster
analysis enabled the identification of various MIT categories. In addition, data was
analyzed on the pattern in the interdependency of variables such as disposition
toward violence, as well as the type and number of crimes. On the basis of the MIT
types identified, numerous measures for investigative police work and prevention
were derived.
Solution components:
 IBM SPSS
What makes it smarter


Early warning triggers and preventative programs put in place
Faster identification of criminal activity and timely prevention
© 2012 IBM Corporation
La Guardia Civil
The need
La Guardia Civil handles large amounts of information every day and used IBM
SPSS predictive analytics software to to analyze data in different fields of
application, such as investigating crimes and psychology. They realized after the
hiring several graduates in this specialty, that standardizing the software into a
single network would improve the department’s performance..
“IBM SPSS software has allowed
us to achieve optimum assignation
of La Guardia Civil’s material and
human resources, and, at the
same time, it has given us the
possibility to establish more
efficient procedures.”
The solution
IBM SPSS was used by multiple departments with different objectives and
functions. For example, the software is used in criminal investigation to support
forensic investigation, enabling analysts to validate expert evidence and apply
scientific methodology to DNA, bullet and handwriting analyses. The psychology
department helps to establish psychological profiles, supports the staff selection
process and helps to develop measures to prevent atypical behavior. The police
force’s department for training studies the public’s security needs and appropriate
training for La Guardia Civil in order to satisfy these needs. In the statistics
department, IBMSPSS software guarantees a real and balanced set of statistical
data, improves the interpretation and knowledge of real world issues that impact
La Guardia Civil and supports and promotes the use of statistics as a
fundamental tool.
-- José Luis Martín Ramos, IT
and Statistics Service (ITSS)
department, Guardia Civil
Solution components:
 IBM SPSS
What makes it smarter

Optimal allocation of resources

Operational efficiency

Consistency and access to same information across departments
© 2012 IBM Corporation
Florida Department of Juvenile Justice
The need
More than 85,000 youth enter this state’s juvenile justice system each year.
FDJJ is dedicated to rehabilitating them and committed to the principle that
proper prevention, intervention and treatment can keep juveniles from
ultimately falling into the adult corrections system. But funding is limited, so
FDJJ must use its resources wisely. To produce the most positive impact on
the communities it serves, the agency needed deep visibility into the inner
workings of the juvenile justice system
The solution
This state government agency, responsible for rehabilitating young offenders,
can't afford to make decisions based on emotion. That's why it uses
sophisticated analytics and statistical models to understand juvenile
delinquency: why it happens, when and where it is likely to occur and how to
prevent it. The agency closely tracks program success, so it can cut or fix
ineffective treatments and focus on what works. Because the success of
individuals is equally important, each juvenile's record is analyzed for risk
factors, compared against statistical models to calculate likelihood of repeat
offenses and then matched with the most appropriate programs. Statistical
models also provide highly accurate forecasts of crime rates across the state,
allowing the agency to plan ahead for resource allocation.
“We are operating in an
environment where funding is
very tightly controlled, so we
need to ensure that the
programs we run have the
greatest possible positive
impact... We can’t afford to
throw money at the problem, so
we have to be a lot smarter
about the way we run our justice
system.”
- Mark Greenwald, Chief of
Research and Planning at
FDJJ
Solution components:
 IBM SPSS
What makes it smarter



Prompted legislative changes that reduced delinquency by 34%
Drove a steady year-over-year decline in juvenile referrals per thousand
people since 1994 by reducing recidivism and crime rates in general
Enables more effective planning and allocation of resources by
uncovering cyclical crime trends; for example, the unintuitive insight that
delinquency rates fall during school vacations
© 2012 IBM Corporation
Mecklenburg-Vorpommern State Police
The need
It is police procedure to record every incident attended, including all the
relevant details. In a large state such as Mecklenburg-Vorpommern, high
volumes of new data are generated every day. In the past, reports paper-based
or done with office productivity software, sometimes resulting in significant
duplication of information. There was no way to aggregate statewide data,
analyze it or make information available to others easily. They realized that to
effectively process inquiries or searches, or conduct tactical resource planning
– a single, cohesive view of information was essential.
“The possibility of combining
information and viewing it from all
angles has opened up completely
new approaches for investigation
work.“
- Thomas Helm, Police Commissioner,
Mecklenburg-Vorpommern State Police
The solution
The organization adopted IBM Cognos Business Intelligence as the foundation
for that system and support the specific objectives for each of several subprojects. IBM Cognos software replaced verbal handovers and printed lists with
a statewide system. All authorized agents are now able to access consistent
and reliable information from all department – analyzing the data to identify
trends or better allocate resources. They are also working on a crime statistics
system that taps into real-time data from the Federal Criminal Police Office.
Solution components:
 IBM Cognos BI
What makes it smarter



Standardizes information and ensures that up-to-date information is
available to 750 decision makers across multiple departments
Provides user-friendly, customizable analysis that helps facilitate everyday
police work plus medium- and long-term financial and personnel planning
Contributes to greater public safety and better, more efficient use of public
funds
© 2012 IBM Corporation
Memphis Police
The need
With traditional policing practices unable to thwart a rising rate of
criminal activity and budgets tight, the Memphis Police Department
pioneered a way to focus their patrol resources more intelligently
The solution
By recognizing crime trends as they are happening, MPD’s predictive
enforcement tool based on IBM SPSS statistics and ESRI AGi gives
precinct commanders the ability to change their tactics and redirect
their patrol resources in a way that both thwarts crimes before they
happen and catches more criminals in the act
“On short notice, we’re able to shift
officers to a particular ward, on a
particular day, right down to the shift
level. It’s a bit like a chess match
and [the IBM SPSS solution] is
enabling us to make arrests we
nevercould have before.”
Larry Godwin, Director, Memphis
Police Department
Solution components:
 IBM SPSS
What makes it smarter
 30% reduction in serious crime overall, including a 36.8% reduction
in crime in one targeted area.
 15% reduction in violent crime
 4x increase in the share of cases solved in the MPD’s Felony
Assault Unit (FAU), from 16 percent to nearly 70 percent
 Overall improvement in the ability to allocate police resource in a
budget-constrained fiscal environment
© 2012 IBM Corporation
Geneva Cantonal Police
The need
The police department had accumulated a large volume of data from a
variety of sources. However, despite updating its software and
databases, the department still struggled to make meaningful use of
the data. Officers found it difficult to locate and access the information
they needed and, once found, the quality of the data was often poor.
The solution
The project was to geocode all the events - such as robberies,
burglaries and assaults - warehoused in the Geneva Police Service
databases. By performing a geocoding operation, the department can
now exploit and analyze the coordinates of a crime. The spatial data
obtained by geocoding and the nonspatial data from the data
warehouse are combined within a report, thus creating a true link
between the business intelligence data and the spatial data.
“This solution not only helps us
target areas where crimes might
occur, but also ensures that
police officers' time is used
effectively.”
—Didier Froidevaux,
Head of Strategic Studies
Department
Solution components:
 IBM SPSS
 IBM Cognos BI
What makes it smarter
 Reduced the number of criminal code offenses by 3 percent
 Gained the ability to predict where and when crimes were most
likely to occur
 Optimized and allocated resources more effectively
© 2012 IBM Corporation
A public safety organization in the Southern United States
Reduced crime rates by 9.2% & number of robberies 31%
The need
Seasonal crime waves were baffling police, who came to realize that crime
patterns were larger than one jurisdiction. But although police could confer across
lines, they still couldn’t drill down into the minute details of other jurisdictions to
take these broader trends into consideration when solving individual crimes. The
department and its neighboring agencies needed to understand trends and
migratory patterns but lacked the ability to easily share data and make
associations that would solve crimes faster using fewer resources.
The solution
This city-county police department is using predictive analytics in completely new
ways by tapping into crime data provided by 38 individual agencies across the
state. The ability to see across boundaries and make informed associations
between crimes and suspects is revolutionizing crime fighting. Local police
officers track crime trends across the state and drill down to individual crimes as
needed to focus resources where they can have the most impact. For example,
the department gained the ability to track specific crimes outside of its jurisdiction
and discovered that local criminals were driving as far as 100 miles to target
tourists along the coast and visitors to NASCAR events held in the state. By
working together with other jurisdictions and taking a larger view of crime, it can
actually pinpoint likely suspects and conditions where extra vigilance has a high
likelihood of success.
What makes it smarter
 Reduced crime rates in the county by 9.2% crimes in one year
 Decreased the number of robberies 31%, from 606 to 418
 Anticipated reductions in crime rates through prevention and higher conviction
rates of suspected criminals
Solution components:
 IBM Cognos BI
 IBM i2
© 2012 IBM Corporation
A metropolitan police department in the United States
Develops an accurate short list of robbery suspects by analyzing past behavior
The need
While crime was decreasing in this large metropolitan region, the cost of fighting
crime was skyrocketing, limiting the effectiveness of the police to tackle the huge
caseloads of reported crimes. Detectives were solving only about 30% of reported
robberies. To solve more crimes, the department needed to narrow down the
possible suspects, a tall order in a city of 2.5 million people.
The solution
This police department is using predictive analytics modeling techniques to catch
criminals faster by narrowing down potential robbery suspects by using their
habits against them. Using data from an existing crime information warehouse,
the analytics solution takes crime classifications, such as time of break-in,
property stolen, method of entry and other facts gathered in the field and then
uses databases defining known criminals, their locations, their way of working and
distinguishing body marks, to identify a list of most likely suspects. For example,
officers in the field can tell the analyst using the tool that a robbery was conducted
during the day using a stolen vehicle in a specific neighborhood, and within
minutes, the analyst can present a list of suspects in the vicinity using matching
methods so that officers can conduct further investigation.
What makes it smarter
 Identified the actual perpetrator as a top five suspect in 18 of 40 randomly
selected solved cases and a top 20 suspect in 29 of 40 cases, a 45 and 73%
hit rate, respectively
 Anticipated a reduction in repeat crimes and property losses through faster
arrests of suspects
 Anticipated lower caseloads for robbery detectives as a result of higher
conviction rates
Solution components:
 IBM SPSS
 IBM BladeCenter
 IBM Storage
 IBM GBS
© 2012 IBM Corporation
A law enforcement agency in Puerto Rico
Lowered its crime rate 20% by using advanced analytics
The need
In recent years, Puerto Rico experienced a significant increase in major crimes
related to drug trafficking, setting records in number of murders, burglaries and
robberies. This law enforcement agency needed to enable its unit commanders to
make more informed decisions and help them establish a more reliable coverage
strategy through better visualization of the agency’s geographical areas, where
they could deploy officers to predict and prevent crime. The agency wanted a
solution that would help it cut the violent crime rate, resolve criminal cases and
reduce the time to arrest criminals.
The solution
This law enforcement agency implemented an advanced analytics solution that
enables it to analyze crime along multiple dimensions using a Crime Information
Warehouse that automatically gathers and integrates information on criminals and
criminal activity from both internal and external sources. Using location based
software to map analyze crime trends, agency officers and analysts can detect
crime patterns as they form, enabling unit commanders to take measures to stay
ahead of trends and prevent spikes in criminal activity. In addition, and unlike
most criminal databases, the CIW integrates key data about known criminals such
as relationships, aliases and tattoos to help the agency visualize a web of
information that can prove to be the key to finding and apprehending criminal
suspects much more quickly than in the past, enabling detectives to spend time
making their case instead of gathering information.
What makes it smarter
 Reduces the country’s crime rate by 20% and helps lower recidivism
 Enables the agency to quickly spot and respond to patterns that indicate an
increase in major crimes
 Helps the agency find & arrest criminals in days or even hours
Solution components:
 IBM Cognos BI
 IBM Power Systems
 IBM Storage
 IBM DB2
 IBM InfoSphere
 IBM GBS
© 2012 IBM Corporation
A crime-prevention unit in Latvia
Reduced the number of unsolved crimes & allocated police resources more effectively
The need
This crime-prevention unit in Latvia was charged with monitoring and preventing
crime in the country, but it struggled with how to piece together information to
crack unsolved cases. The unit sought a solution that used a central data
warehouse to combine data from existing crime databases across the country.
The solution needed to be capable of selecting and analyzing this data in complex
ways, yet simple enough to enable staff without specific IT skills to compile and
share detailed reports. The solution should also support investigations with both
operational and analytical power - for example in-depth searches for people,
objects or crimes across the country - and do it in as close to real time as
possible.
The solution
With the help of an IBM Business Partner, this crime-prevention unit in Latvia
uses business intelligence to gather crime information into one place and analyze
it in near-real time, enabling authorities to see crime patterns as they form. For
example, agents may see that car thefts focus on a certain make of luxury car, or
that these thefts are closely followed by a burglary. Instead of trying to catch up
with criminals, police can anticipate where they will strike next and use new data
to solve old crimes. With predictive analytics, crime pays less than ever.
What makes it smarter
 Facilitated a more proactive approach to crime solving, with the ability to see
crime trends as they happen
 Enabled a more efficient use of technology and human resources
 Reduced the number of unsolved cases and enabled authorities to solve
crimes more quickly
 Decreased incidents of data and processing errors with automated workflow
and analysis
Solution components:
 IBM Cognos BI
© 2012 IBM Corporation
A metropolitan police department in the United States
Projected 99% improvement in connecting officers with evidence and leads
The need
Crime is dropping in the United States. From advanced forensics and biometrics
to sophisticated analytics solutions, police departments across the country are
looking at evidence a whole new way and pooling resources with other agencies
nationwide to predict a criminal’s next move. This metropolitan police department
knew there were insights to be unlocked from its large pool of criminal data, but it
lacked the tools to analyze the data in a timely manner to act quickly.
The solution
A robbery on one side of town and a speeding ticket on the other side of town
may not be related. However, if they are, they could lead police to a pattern of
incidents that can help catch a criminal. This police department in the western
United States wanted to be able to connect these dots to solve crimes more
quickly. It implemented a criminal data analytics solution that takes crime
information from several databases, analyzes it and maps it on a highly visual
interface that lets officers easily spot, for example, whether heightened gang
activity usually occurs before a certain type of robbery, or whether car theft occurs
afterward. Not only can the department draw these connections, but it can also
link to dozens of law enforcement agencies in neighboring states to compare
evidence and potentially catch suspects on the run.
What makes it smarter
 Provided the capability to share information and collaborate with more than
3,000 law enforcement agencies across the United States
 Projected to connect officers with evidence and leads within minutes instead of
days, weeks or longer, a more than 99% improvement
 Expected to lead to more crimes solved with the ability to make non-obvious
connections and correlate vast amounts of data from multiple disparate
repositories
Solution components:
 IBM Cognos BI
 IBM i2
 IBM GBS
© 2012 IBM Corporation
Download