Data Warehouse – National State and Local Perspective Presenters: Mark Takayesu, Orange County Department of Child Support Services, Mtakayesu@css.ocgov.com Tank Waklee, California Department of Child Support Services, Tank.Waklee@dcss.ca.gov Jennifer Effie, San Diego Department of Child Support Services, Jennifer.Effie@sdcounty.ca.gov Data Warehouse – National State and Local Perspective What Is A Data Warehouse? While states vary in definitions – A Data Warehouse is a Computer System that stores information to facilitate reporting and analyzing data to obtain information and facilitate decision making Source – “Enhancing Child Support Efforts: Summary of Data Warehouse Efforts in Nine States” – The Lewin Group 2006 Data Warehouse – National State and Local Perspective Want to cook Pumpkin Ravioli with Sage Brown Butter for an important client for dinner? Data Warehouse – National State and Local Perspective Data Warehouse-National State and Local Perspective Data Warehouse – National and State Perspective • Data Modeling Workgroup – Sponsored by the Federal Office of Child Support Enforcement – Facilitated by Richard Ordowich • Began Meeting in March of 2009 • Objective -Develop a standard data warehouse model and data definitions, and provide a forum for data warehouse sharing between states • Have a deliverable within 12 months • Developed based on business need Data Warehouse – National and State Perspective • Participating States and Organizations – 14 states •There is no “one size fits all” data warehouse OCSE ACF Vermont Massachusetts Illinois Virginia Colorado Texas Hawaii Washington Louisiana Minnesota California Arizona North Dakota Nevada Data Warehouse – National and State Perspective Washington Policy development research and statistical analysis: help staff understand cause and effect analysis (how case workers actions affect outcomes). Cost Avoidance Studies Illinois Ad-Hoc reports for managers to filter through cases requiring action. Monthly Management Dashboard Reporting Minnesota Operational Research Analysis, Fraud Investigation, Quality Assurance Review Hawaii Business Process Management (streamline staff’s work processes, customer service etc.) Quality Assurance, Performance Measurement Vermont Data Mining – medical support (focus on cases with highest potential for medical support), audit prediction, payment prediction need for intervention, operational reports for supervisors, Dashboard, Ad Hoc reporting Data Warehouse – Vermont Example Data Warehouse – Vermont Example Data Warehouse – Vermont Example Data Warehouse – Vermont Example Data Warehouse – Vermont Example Data Warehouse – Vermont Example Data Warehouse – Vermont Example Data Warehouse – Project Objectives • Agree to start with the OCSE – 157 • Develop a model based on –Dimensions –Facts Data Warehouse – Project Objectives Examples of Dimensions And Facts Dimension Examples (descriptive values) Time Year, Month, Day, etc Case Info Interjurisdictional Code (International/Tribal/Interstate/No Jurisdiction) IV-D/Non IV-D Code Interjurisdictional Status (Initiating/Responding) Person Info Person Role Code (CP/NCP/Child/Employee) Medical Coverage Code (Y/N) Data Warehouse – Project Objectives Examples of Dimensions And Facts Fact Examples (numeric values used in calculations) Case counts Person counts Current support due at month end $ Current support distributed at month end $ Total arrears due at month end $ Arrears distributed at month end $ Data Warehouse – Project Objectives Line Identifier Line 2 Line 2a Line 2b Description Cases WithSupport Orders Established Interstate Cases Initiated Interstate Cases Received Ar re ar s Cu r Su rent pp or t ed Ob lig at Pa re nt Fact Areas ag e OCSE 157 Annual Report Fact Descriptions Data Warehouse – Orange County Example Interstate Initiating vs. Responding Monthly Percent of Current Support Collected Initiating = 8,345 Cases, Responding = 3,807 Cases 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% Initiating 55.7% 44.3% Oct-08 60.5% 51.4% 39.4% Nov-08 45.0% Dec-08 Responding 55.1% 54.9% 40.4% 40.3% Jan-09 Feb-09 56.9% 58.8% 43.7% Mar-09 55.0% 55.2% 42.9% 40.0% 42.8% Apr-09 May-09 Jun-09 Data Warehouse – Orange County Example Interstate Initiating Monthly Percent of Current Support Collected Current, Former Never 80.0% 60.0% 40.0% 20.0% 0.0% Oct-08 Nov-08 Dec-08 Jan-09 Feb-09 M ar-09 Apr-09 M ay-09 Jun-09 Current 26.7% 28.1% 28.4% 28.7% 24.0% 27.9% 24.2% 22.3% 27.1% Former 44.3% 40.5% 46.3% 40.3% 42.7% 47.4% 47.2% 43.0% 46.9% Never 48.9% 41.5% 48.4% 43.6% 42.6% 44.8% 44.4% 42.2% 43.4% CSE BDDR • • • • • • Why What Where Who How And then Some Why • Why we Got Here! – Reports requests take too long – Query requests take too long – Not enough State Staff resources to fulfill these requests – Listening to you (Top 5) – This is YOUR System/YOUR Data What • What is BDDR – “Back Door Data Repository” – Raw Data from CSE – De-normalized where possible • Based on Counties perspective – Name competition What BDDR is Not • Buzz Word - Data Warehouse • Core data is not – Aggregated – Summarized – Some History What Software • How can I access BDDR – MS Access – SQL Server Mgmt Studio – Toad – SAS – Fox Pro – Anything that allows port 50217 What’s in BDDR Now • • • • • Approx 75 Tables Core CSE Data 1257 34/35 Misc – – Dax Reports – EI – CSLN • What ever data you want! - - Did I just say that!!! What Means What What’s in BDDR • Types – Append, Complete, Snapshot Where • BDDR is a 2005 SQL Server located in Rancho Cordova OTECH Data Center • Uses the DTS Network (not internet) • SQL Port 50217 Who • Who is Using BDDR How • How do I get access • Two users per county • How does the data get there eBHC OTECH DCSS Project Office Future Plans Near • More Data – Employer Demographics – Task – Guideline Calculator – What ever the Counties Request • Data Dictionary • Reporting Services • BDDR Blog Future Plans Near Future Plans Far • • • • AIX DB2 Fully Automated ETL Faster loads from CSE – Bi-weekly – Weekly – Daily Future Now System Performance Data Analysis • Quantitative vs. Qualitative Evaluation – Quantitative: 1257 line comparisons between counties, months, years, etc. – Qualitative: Why are the case counts and dollars moving up or moving down? – the goal is to discover the story behind the numbers…that’s what BDDR gives us Access to Data • • • • • Identifies problems Maximizes resources Communicates project results Enhances decision making ability Allows for continuous program improvement Pieces of a Puzzle • Querying & analyzing data is part science and part creative writing – Success requires time, patience, & vision – Collaboration among IT, Analysts, Child Support Staff, other counties, & “puzzle-solvers” Information Sharing • Sharing information and resources is key to our success – BDDR Sharing Forum & PAW Blog – Data Dictionary – OC & SD Documentation – Case Management Tool (Los Angeles) – County SMEs San Diego Documentation 41 Data Groupings • • • • • • • Case Participant Legal Accounts/Payments Public Assistance 34/35 and 1257 Reference Info Case Data • • • • • • • • • Case and Account Summary Case Status History Case Type History Case Management History Case Assignment History Interstate Perspective Events Review & Adjustment Early Intervention Activity Logs Participant Data • • • • • • • • • • • Case Participant Participant Relationship Participant Demographics Demographic Source History Phone Numbers Prisoner Information Income Employer Wage Assignment License Liens Legal Activity • Court Case • Legal Activity (including BK, WC, Contempt, Liens, and Support) • Service Activity • Support Terms Accounts & Payments • Case Accounts • Account Transactions and Summary Transactions • Charging Instructions • Logical & Physical Collections • Remittance Advice • Disbursement • Payables Public Assistance • • • • Participant Public Assistance Public Assistance Case Public Assistance Case Balance Public Assistance Case Transactions CS 34 & 1257 • CS 34/35 – Collections – Disbursements – Undistributed Collections • 1257 – FFY Detail for 2006/2007, 2007/2008, and 2008/2009 – Financial Detail – FFY Summary Building a Query Assessing MNO Child Only Cases using Microsoft Query Tool 49 Building a Query 50 Building a Query 51 Building a Query 52 Building a Query 53 Building a Query 54 Building a Query 55 Parting thoughts… • • • • • Find a tool that works for you Don’t reinvent the wheel Ask questions Share Ideas Study, study, study!!! Thank you Questions? (Besides why Tank is named Tank”!)