Data Mining: Unlocking the Intelligence in Your Data Marlon B. Williams, CPA, ACDA Partner, IT Advisory Services Weaver 0
Today s Agenda Big Data What is it? Data Mining at a Glance Why the Accounting Profession needs Data Mining Know Your Tools Next Steps & Final Thought Questions & Answers 1
BIG DATA What is it?
What is Big Data Data is produced from all parts of a company Wikipedia says: A collection of data sets so large that it becomes difficult to process using traditional data processing applications. Marlon says: So much data that it freezes Microsoft Excel. 3
What is Big Data A company generates multiple sources of data It s not uncommon for each data source to have 50,000+ lines 4
What is Big Data Having large amounts of raw data does not necessarily translate to effective decisionmaking Interpretation is key, and this is where data mining shines 5
Data Mining at a Glance Start with large amounts of disparate, raw data from different sources Vendor Data Payroll Reports HR Reports Process and interpret the data using: Software ACL Scripts and formulas Industry statistics Artificial intelligence Database management Accounting knowledge Output is usable data: Reveals correlations and patterns to support decision making 6
Data Mining at a Glance Manual Sampling See only part of the picture Eyeball large reports Make decision based on small parts of the data Constrained by time Automated See the full picture Leverage all of the data Combine data from all parts of the company Better data in less time Analytics that are customized to your processes 7
Application to various industries: Prime Candidates for Data Mining Disparate systems Manual processes involving data entry Large Databases Industries with trends Folks that want to understand the information they collect 8
: Prime Candidate for Data Mining Payroll Compare employee to vendor listing. Ghost Employees PO box matches Review proper cutoff Social security matches Employees without Company e-mails Duplicate Employees Credit Card review Improper merchant accounts Activity by employee Split Transactions Duplicate transactions Unusual activity 9
: Prime Candidate for Data Mining Vendor Review Top vendors Vendors in your system without recent charges Vendors with PO boxes close to employees homes Consistency of vendor master file Duplicate vendors with similar names User Access Excessive invoices General Ledger review Entries without descriptions Key word searches Approval by same party Duplicate transactions Debits to revenue accounts Credits to Cash accounts Large manual entries Activity in Dormant accounts Even dollar entries 10
Oil & Gas Industry: Prime Candidate for Data Mining JIB Reporting: Sample Cost Questions Well Costs Do costs to production vary from expectations? Vendor Costs Do vendors charge for the same price for similar services? Have new vendors been properly vetted? Production trend analysis by various drivers: cost to barrel, MCF by well, region, location, project, etc. Looking at trends within production against geologists reports; pinpoint nonproductive elements 11
Oil & Gas Industry: Prime Candidate for Data Mining JIB Reporting: Sample Cost Questions Transportation Costs Analyze trends within the costs to better explain truck tickets Is it cost effective to develop your own disposal well? Production Costs Volume vs. sales are there discrepancies that may point to a significant shrinkage problem that needs to be investigated? Timing of production can we maximize efficiencies? Should we convert from electric compressors to natural gas? 12
Oil & Gas Industry: Prime Candidate for Data Mining Outliers Reveal Patterns 13
Know Your Tools Excel vs. Other Tools Excel is good to a point Maxes out, freezes up One changed formula has domino effect Data mining tools Made for very large data sets Sophisticated data analysis Custom queries that are repeatable and reliable 14
Trucking Expense 15
Next Steps Know your data If you don t know your data, you will not develop a successful program Key points to consider in designing a successful program: Develop a plan Identify the risks Determine your stakeholders Determine the tool and who will implement the tool 16
Final thought: Mining your company s data can unlock key information to reduce risks and improve your bottom line 17
Questions & Answers Marlon B. Williams, CPA, ACDA Partner, IT Advisory Services Weaver Direct: 972.448.6919 Marlon.Williams@Weaver.com 18