Putting the ‘science’ into data analysis
Traditionally, financial planning and analysis have been about financial data. Assessing performance against budget has been a cornerstone of financial management. And it’s relatively straightforward: the mathematics involved is no more complicated than the calculation of percentages.
On the other hand, understanding the causes of performance or divergence from budget requires a better understanding of the business. It means asking pertinent questions and carrying out root cause analysis. And higher-value analysis, such as predictive or prescriptive analysis, demands more advanced analytics that usually requires the expertise of a data scientist (Figure 3).
Figure 3 The Gartner Analytic Ascendency Model and Finance Function Reporting Focus xviii
Framing the problem: Clarify the actual issue: translate an unclear request into an analytical problem.
Data collection Determine what data is available and might be useful; ask what other data could be accessed.
Data cleansing Correct errors and missing, incomplete or corrupt records; convert to useable form. Data exploration Look for trends, maybe with visualisation and regression analysis to identify correlations. Data analytics Data mining, statistical modelling, machine learning, deriving algorithms. Communication Visualisation and storytelling to help users understand and action findings.
“Higher-value analysis, such as predictive or prescriptive analysis, demands more advanced analytics that usually requires the expertise of a data scientist.”
Data — Deluge and Decisions | Episode 1 | Welcome with Andrew Harding
Readying business for the Big Data revolution — Russell Goldsmith is joined by the Association's Peter Simons and Simon Jeffery at Siem Car Carriers
