T-22 Predictive Analytics in Action-San Diego

Transcription

T-22 Predictive Analytics in Action-San Diego
Predictive
Analytics in Action
San Diego County
San Diego County
•
•
•
•
•
4,083 sq. miles – 8th largest county by size
3 million residents – 5th largest by population
725,000 children
13.9% living below poverty level
Unemployment rate 5.9% (as of 10/14)
County of San Diego
Department of Child Support Services
Our mission is to enhance our children’s futures by obtaining support for
families today
•
73,000 cases
•
Over 113,000 children served
•
FFY 2013 - $178.7M
•
6% of California’s total caseload
•
FPM Performance FFY14 : #1 – 99.8%, #2 – 88.2%,
#3 – 70%, #4 – 69.8%, #5 - $3.85 (FFY2013)
Predictive Analytics
•
Precision
•
Consistency
•
Agility
•
Speed
•
Cost
Overview
Child Support Payment Predictor


Scope of Study:
 1,033 cases opened 10/2011 through 5/2013
 No prior support order
Objective:
 To establish a “payer” classification predictive tool that
prognosticates the NCPs who are potentially at risk to fall
behind on their child support payments.
 Provide staff with a tool to more effectively perform
intervention on cases.
Child Support Payment Predictor

Data Elements
 466 data elements from the case and participants
 68 field variables were created as candidates for
exploratory data analysis
 29 variables were selected for model building
 8 variables remain in the final model
Child Support Payment Predictor

Model data points
 NCP’s Employment stability
 Incarceration history
 Unemployment in previous year
 Living status in previous year
 Default history on first order
 NCP’s email information
 NCP’s employment status at initial order establishment
 Other NCP Demographic data
Child Support Payment Predictor

Payment Predictor Details
 Multinomial logistics regression
used
 4 payer groups created:
 Extremely rare payer (0%-30%)
 Rare payer (31% to 50%)
 Attempt payer (51 to 80%)
 Constant payer (81% or greater)
Distribution of 6 month average payment rate
(N=1033)
Child Support Payment Predictor

By payment class
 Extremely rare payor (0%-30%)
 Rare payor (31% to 50%)
 Attempt payor (51 to 80%)
 Constant payor (81% or greater)
Total due vs. Total payment for each payor
class
Child Support Payment Predictor

Time to order or “waiting time”


Shortest - 2months
Longest- 19 months
Child Support Payment Predictor

Variable – NCP e-mail address


54% had e-mail addresses
Strong impact on payer class
Child Support Payment Predictor

Results:
 Prediction accuracy of the model
 62.5%
 Percentage of the caseload more accurately monitored
 14%
 Able to target/focus on at risk cases before they
become delinquent
What’s Next or in the works?
Business Intelligence – “Smarter Planning”
 New analytics tools
 Greater flexibility and detailed analysis
 Review “payment predictor” for areas of further
analysis
 Legal Paperless System – customer behavior
 Performance-based dashboard
 Additional data point research

A Smarter WAY TO PLAN
Overview


What is Business Intelligence?
 Set of theories, methodologies, processes and technologies that
transform raw data into meaningful and useful information.
Why do we need Business Intelligence?
 Reaching limit of producing NEW reporting and analytical results.
 Paradigm shift from reactive reporting (the past) to proactive
reporting (looking into the future).
 Who is likely to participate in the process?
 Who is likely to pay their child support?
 What will our caseload look like in 5 years?
Smarter Planning – A new way to plan



Smarter Strategic Planning
 Strategic Planning Tool – predict future case loads, staffing needs, budget and
performance.
Smarter Case/Knowledge Worker
 Case Stratification Tool - Based on certain case data predictors, will be able to
predict case outcomes with high probability.
 “Reduce” Case Loads: Focus on cases that need ‘personal touch’ – automate
others.
Change Individual Case and Family Outcomes
 Proactive Case Management
 Targeted Early Intervention
 Keep Families Intact
Smarter Planning – How do we get there?

4 Phases
 Phase I: DCSS Executive Dashboard

Phase II: DCSS Smarter Planning

Phase III: County-wide Data Sharing

Phase IV: County Smarter Government
Smarter Planning – Phase I

Phase I: DCSS Executive Dashboard
 The Past and Present
 Data Visualization of Current Performance Measures
using Key Performance Indicators (KPI)*
 Summary of Case Load Performance
 General Health of Department
 Central Reporting Portal
*A set of quantifiable measures that a company or industry uses to gauge or compare performance in
terms of meeting their strategic and operational goals.
Smarter Planning – Phase II

Phase II: DCSS Data Mining and Predictive Analytics
 Looking at the Future
 Predict future case performance, participant and staff
behavior
 Correlate Staff Demographics to Case Data
 Smarter Planning Tool
 Case Stratification Tool
 Alerts
Smarter Planning – Phase III

Phase III: County-wide Data Sharing
 Current limits to Business Intelligence based on only
Child Support Data (silo effect)
 Partner with other Departments including IV-A to Pilot
 Richer Business Intelligence and Predictive Analytics
Smarter Planning – Phase IV

Phase IV: Smarter Governing
 Executive Dashboard, KPIs and Reporting Portal Rollup for PSG
Departments
 Smarter, Integrated Knowledge Workers
 Probation Officer helping with Child Support Case
Management
 Early Case Intervention
 Reduce recurring public sector demands
 Central Case Management Customer Service Portal
 Predicting Future Budget, Facilities and Staffing
Credit Score
FICO.com
DRAFT Executive Dashboard
Predictive Analytics – Google Now
Questions?
Jeff Grissom
Director
San Diego County
[email protected]