Archive for November 20, 2006

Identifying Gaps in Data Sequences

proc gap is a command language procedure to identify missing numbers within sequences, such as check numbers, invoice numbers, etc. This facilitates the commonly used audit test to identify gaps in data sequences.

Dollar Unit Sampling

Proc SAMPLE is a command language procedure to provide audit population sampling procedures. Both cumulative dollar amount (cma) which is also known as dollar unit sampling (dus) as well as random and interval sampling are supported. The auditor may provide the random numbers, interval values, R factor and J, etc. Description of algorithm used.

Identifying Outliers with Proc Outlier

Identification of outliers in data collections through use of scripts and GUI. This is a command language script which can identify those transactions which are outside what would be expected from a population which is normally distributed. This is a common audit test.  View example screen shots.

Spike Analysis - Analytical Review of Trend Lines

The identification of unusual (upwards) spikes during a timeline is often a part of the analytical review procedures to ensure that the data is reasonable. This command language procedure allows the analyst to identify and focus on those transaction. The auditor can provide optional parameters include the minimum number of transactions required in a sequence, as well as the minimum amount of the fluctuation.  See example screen shots.

Duplicate Payments

Proc RV is a command language procedure to facilitate cash recovery. Cash recovery from supplier overpayments is often possible regardless of how well the internal controls have been designed. Best practices and stringent controls can reduce duplicate invoice payments to less than half a percent. However, this error rate seems largely unaffected by automation. There are a variety of reasons why this is the case - multiple (non-integrated or non-communicating) systems, human error, intentional override, system failures, etc. This is a commonly performed audit test.  See example screen shot.

Producing Histograms using Proc Histo

proc histo is a command language procedure to compute data to be used for histograms. The input file must be in the industry standard tab separated value (tsv) format. Parameters, provided using a script, determine how the elements of the histogram will be computed. The output from the audit procedure is a tab separated file which may be imported into MS-Excel for charting. View script/sample output.

Sorting Data using Proc Sort

Proc SORT is a command language procedure to sort data in order that other analytical procedures can be performed using the BY statement. Although intended primarily for use as a command language, it can also be used with the GUI provided. Examples of other audit procedures which use the by statement include means, benford, outlier, histo, rv and summary. Note that the data to be sorted must be in standard tab separated value format. Up to five summary sort fields may be specified. See example screen shot.

Summarizing Data using Proc Means

PROC MEANS is a command language procedure to provide the basic audit statistics (mean, minimum, maximum etc) for a population in standard tab separated value format. Further details.

Example Results (Charts) from Benford’s Law

Show charts of results from the application of Benford’s law to a set of test data. There are six audit tests which can be performed: F1- first digit, F2- First two digits, F3 - first three digits, L1 - last digit, L2 - last two digits, D2 - second digit

Testing the Application of Benford’s Law

Proc BENFORD is a command language procedure (and GUI) to test the conformity of a distribution with Benford’s law. The results of this audit test are assessed using Kolmogorov-Smirnov in order to provide a measure of the degree of fit between observed data and expected data.  View examples.