Jennifer Robinson SAS Steps to Datadriven Decisions Create a shared vision Provide visualizations and reports PROACTIVE INFORMATIONAL ANALYTIC ID: 586989
Download Presentation The PPT/PDF document "Using Analytics to Unravel Complex Socia..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Using Analytics to Unravel Complex Social Issues
Jennifer Robinson
SASSlide2
Steps to Data-driven Decisions
Create a shared vision
Provide visualizations and reportsSlide3
PROACTIVE
INFORMATIONAL
ANALYTIC
MATURITY
Standard
Reporting
Ad Hoc
Query and
Reporting
Query
Drilldown
Alerting
Statistical
Analysis
Forecasting
Predictive
Modeling
Optimization
What is the best outcome?
Decision-Making SpectrumSlide4
PROACTIVE
INFORMATIONAL
ANALYTIC
MATURITY
Standard
Reporting
Ad Hoc
Query and
Reporting
Query
Drilldown
Alerting
Statistical
Analysis
Forecasting
Predictive
Modeling
Optimization
How many people without homes are incarcerated each year?
How many male
inmates have used EMS twice in six months?
Which
inmates have been treated for addiction?
Which
inmates should be matched with support services?
What is the best outcome?
Decision-Making SpectrumSlide5
PROACTIVE
INFORMATIONAL
ANALYTIC
MATURITY
Standard
Reporting
Ad Hoc
Query and
Reporting
Query
Drilldown
Alerting
Statistical
Analysis
Forecasting
Predictive
Modeling
Optimization
How many people without homes are incarcerated each year?
How many male
inmates have used EMS twice in six months?
Which
inmates have been treated for addiction?
Which
inmates should be matched with support services?
Who is at risk of
recidivism?
What
is the financial impact of an increasing homelessness population on our jails?
What impact could we expect from
permanent housing and support services?
What is the best outcome?
Decision-Making Spectrum
What is the next best action for
preventing incarceration of homeless males?Slide6
Unraveling homelessness
Provide vulnerable people with the right services
Problem
Demand for mental health and substance abuse treatment, care coordination services, and supportive housing far outstrips the supply.
A
highly vulnerable segment of the population repeatedly cycles through local jails, emergency departments, homeless shelters and other public systems.
ObjectivesProvide decision makers insight into the current status of homelessness in their jurisdictionsUtilize data and analytics to assess different facets of the situationLeverage data and analytics to make best decisionsChallenges
A large number of entities, government and non-profit, provide services to this population (housing authorities, health services, law enforcement, etc.)
Obtaining clean data and matching clients across service providers is very difficultSlide7
Unraveling homelessness
Approach for utilizing
analytics
What is the profile of high utilizers of hospitals and jails?
What are key factors that correlate with high utilizers?
What
support services can stop the cycle of high utilization?
How can support services be delivered to targeted population?
Are
the shelters and housing
adequate and in
the right place?
What locations should additional resources go to?
How to bend cost curve for top utilizers of services?
How to prevent homelessness for vulnerable subpopulations?Slide8
Unraveling homelessnessSlide9
Unraveling Homelessness
Phase 1Slide10
Post-conviction Criminal Justice Data
Unraveling Homelessness
Data is the foundation of analytics
SAS® Data Management Layer
SAS
®
Analytics
Homeless Counts
HMIS Data
Shelter Services Data
Subsidized Housing Data
Supportive Services Data
Social Services Benefits
Emergency Room Services or EMS
Arrest Record History
Incarceration History
911 Data
Individual Characteristics
Homeless Data
Homeless Services Data
Social/Health Services Data
Imprisonment History
Parole History
Incarceration History
Court Record History
EMS Records
Emergency Room Services
Social Services Benefits
Mental Health Services
Mental Health Benefits
Pre-conviction Criminal Justice DataSlide11
Unraveling Homelessness
Phase 2Slide12
Post-conviction Criminal Justice Data
Unraveling Homelessness
Data is the foundation of analytics
SAS® Data Management Layer
SAS
®
Analytics
Homeless Counts
HMIS Data
Shelter Services Data
Subsidized Housing Data
Supportive Services Data
Social Services Benefits
Emergency Room Services or EMS
Arrest Record History
Incarceration History
911 Data
Individual Characteristics
Homeless Data
Homeless Services Data
Social/Health Services Data
Pre-conviction Criminal Justice Data
Imprisonment History
Parole History
Incarceration History
Court Record History
EMS Records
Emergency Room Services
Social Services Benefits
Mental Health Services
Mental Health BenefitsSlide13
Unraveling child abuse
Identify who is most at risk - at the right time
Intake Screening – assess risk based on initial reports
Provide risk scores
Create networks and timelines
R
esolve entity issuesOngoing case management – alerts generated on open CPS casesGenerate automated alerts for open casesSurface ancillary data relevant to cases
Gain insight from complex dataSlide14
Unraveling child abuse
A caseworker
can be confronted with a huge volume of information
.
Translate information from disparate data
sources into a risk score summary.Slide15
Unraveling child abuse
A caseworker
can be confronted with a huge volume of information
.
Translate information from disparate data
sources into a risk score summary.Slide16
Unraveling child abuse
Summarize years of CPS events for individuals related to a case.Slide17
Unraveling child abuse
Relationship map without entity resolutionSlide18
Unraveling child abuse
Relationship map with entity resolutionSlide19
Unraveling child abuse
Many
systems
suffer from data gaps caused by individuals with multiple IDs within and between CPS jurisdictions. The
Visual Investigator tool resolves these gaps to yield a more complete report history for alleged
perpetrators and victims.Slide20
Unraveling child abuse
New information
changes the risk score.
Therefore, alerts are sent to case
workers when new information pertinent to a case becomes available. Slide21
Unraveling opiate abuse
Identify prescribers, dispensers, and patients
Prescribers
Which prescribers have suspicious or unusual behavior patterns that warrant further investigation?
What are the common characteristics of prescribers who engage in inappropriate and/or unlawful prescribing patterns?
Dispensers
Which dispensers are not reporting filled prescriptions accurately and/or timely?
Which dispensers may be viewing patient prescription records inappropriately?
Patients
Which patients have suspicious or unusual prescription patterns that warrant further investigation or intervention?
Which patients may be intentionally misrepresenting identity and/or the victims of identity theft, undertaken for the purpose of obtaining controlled substances inappropriately?Slide22
Unraveling opiate abuse
Data
Collection
Allows reporters to report prescription data on a daily
basisEmploys data from numerous external sources
Robust identity resolution using probabilistic matchingAnalyticsContinuous analytics to produce alerts and external reporting dailyEnables advanced geospatial analysisUses advanced analytics to detect emerging schemes and hard-to-find behaviors
Workload ManagementProvides alert triage for automated workflowAutomated external ReportingSlide23
Unraveling opiate abuse
Alert
scoring with comparison to dispenser peer
groups within a compressed network.Slide24
Unraveling opiate abuse
Alert
scoring with comparison to dispenser peer
groups within a compressed network.Slide25
Unraveling opiate abuse
Alert
scoring with comparison to dispenser peer
groups within a compressed network.Slide26
Unraveling Student Success
Assist teachers and students to achieve their best
Growth models based on reflective and forward looking analytics
Reflective
Student Growth Metrics
Forward-Looking
Student ProjectionsEducators
How much have teachers/schools/districts influenced student growth in the past?
Students
How likely are students to be successful in the future?Slide27
Unraveling Student Success
Assist teachers and students to achieve their best
Reflective
Student Growth Metrics
Forward-Looking
Student ProjectionsElements
Growth models based on reflective and forward looking analytics
Accurate linkages between teachers and students to attribute the amount of growth a student has experienced with each teacher
Results as a serviceReports that dig into diagnostic information as well as customized reports so that administrators can explore dataSlide28
Correlation
Achievement
and PovertySlide29
Correlation
Academic growth
and PovertySlide30
Using Data Differently
Be proactive vs. reactive
Identify
and prioritize highest-impact issues
Combine data from many data sources to improve informed decision making
Seek actionable insights from dataUse data to inform meaningful policy, procedure, and practice.
If you
c
hange
the way
y
ou
look
at things,
t
he
things
you look at
c
hange.
- Wayne DyerSlide31
Thank You
Questions? Jennifer.Robinson@sas.com