Employee Compliance Case Study: Model, Report & Query Design Mark Wilcock Zomalex Limited
Agenda
Mark Wilcock • Independent consultant and trainer in data analysis, reporting and visualisation • Background in investment banking • Microsoft MVP, Data Platform • Organiser of community meetup London Business Analytics Group • MA (Natural Sciences, Cambridge) • MBA Imperial College London • mark. wilcock@zomalex. co. uk
Setting the scene
Employee activities of interest to a bank’s Compliance function • Pre-employment Screening • Mandatory Training • Personal Account Dealing (for all accounts in employee’s family) • Outside Appointments • Private Investments • Gifts & Entertainment • Breaches Each of these for a different source system
What’s special about this data? 1. It’s small. 2. It’s observational. 3. Have a think – let’s discuss later.
Demo
Demo – A single view of the employee
Five easy steps
Making it happen - in 5 easy steps(!) 1. Connect - to several disparate systems and import the data 2. Clean – consistency, quality, integrity, . . 3. Model – the spider's web of employee activities to get holistic view 4. Visualise - communicate to the audience 5. Publish - share and distribute
1. Connect
2. Clean
3. Model
4. Visualise
5. Publish
Power BI provides a single view of the customer • Connect to most data sources • Clean, prepare and transform data • Model the complex connections between the employee and various activities • Visualise, drill, interact, slice and navigate • Publish & share insight (dashboards, Q&A, insights, mobile)
Final Remarks
What’s special about this data? 1. It’s small. 2. It’s observational. 3. Have a think – let’s discuss later.
Quis custodiet ipsos custodes?
A Hippocratic oath for data scientists? This file comes from Wellcome Images, a website operated by Wellcome Trust, a global charitable foundation based in the United Kingdom
Thank you • mark. wilcock@zomalex. co. uk London Business Analytics Group • March 27 th : Panel discussion • April 3 rd: Power BI Training • April 11 th : Qual vs Quant