Establishing a business modelfor better, faster, cheaper stats Presentation to SUNZ conference 21 February 2013 Wellington Gavin Knight
Better Faster Cheaper StatsWhy? – Business driversHow? – Copy VicPol and WAPol – Business model and technologyWhat? – Technology – Analytical Data-martsWhen? – What has been built – What is being built – What might we do nextWhat have we learnt?
Business drivers Cost and speed to market – Statistical development cycle vs SDLC • more BAU and less projects Less reliance on analysts in districts – Eliminate duplication – Best use of different skill levels Centralised single source of the truth Advanced analytics – RIRO partly mitigated – Predictive analysis
Case Case BDM File hierarchy-view Occurrence OffencePerson Person
Case BDM process-view Response process Investigations process Initial Call for attendance service Acceptance Dispatch/Assignment Arrival Police Police Offence Solved complete discover discover offence offence absent with offender offender Pending PendingReceive Initial Active Assess report dispatch/ physical Link offender report attendance investigation assignment attendance
Development plan Done – SAS DI – Knowledge transfer – DR Environment – CASE Data-mart – Justice Sector KPIs Doing – Applying analytics (matching, imputation, estimation, merging) – Business processes / SLA – FV/ODARA – Repeats – BI strategy Likely Future – Predictive analysis – Roll out more standard reports – SAS EBI – Case Management – Geo-spatial – Replace INCOFF with RCVS and more
What have we learnt? Technical-rational – Brilliant! Institutional – Values Implications – This type of change is slow – It is a catalyst for broader change – Vision - commitment – competence – High risk – high reward