Try IDEA

Back to Basics: Forecast Into The Future With Trend Analysis

Back to Basics highlights foundational IDEA skills every auditor needs in their toolbox. Calculate trends and forecast values into the future with the technique Trend Analysis.  Use this task to easily see whether the data contains an upward or downward trend. Let’s get started…

 

You can use Trend Analysis to the trend of data over time and forecast values into the future.  Using this type of technical analysis is important in finding and confirming trends.

 

After running Trend Analysis, you will be presented with a graphic that contains:

 

    1. A light blue line representing Actual Data
    2. A light orange line representing Reference Data
    3. A green line representing Trend Line/Fit
    4. A red line representing Forecasts

The Trend Analysis task in IDEA adds the trend line, which is an easily recognizable line that connects a series of data points. However, IDEA does more than place the trend line. With IDEA you can:

    • Specify how many periods into the future IDEA should forecast
    • Generate forecasts on a group-by-group basis in the database (audit units that refer to a group of data including location, division, franchisee, branch, or other reporting entity)

Of course, the forecasts are only valid if you expect the past trend to continue.

Trend Analysis results are then exported to databases or a Results output. From the Results output, you can then view the results of the Trend Analysis in a grid or a graph.

Viewing the Results output in a graph allows you to see if the data contains an upward or downward trend. You can also discern how reliable the predicted values are based on where actual data, fitted data, and forecast data fit on the graph.

The trend line and the actual (original) data will be plotted in the Results output.

The coefficients of the trend line, the differences between the values of real data and the values on the trend line, and the forecasting values (when requested) are calculated and can be written to output databases. The task also produces a MAPE score that measures how well the trend line fits the actual data. You can have more confidence when using data with low MAPE values as this means that the past data has a good fit to a straight line pattern.

A Real Life Example of Using the Trend Analysis Function

You can use the Trend Analysis task to predict the Marketing Cost for the upcoming quarter.

Here Marketing Cost will be a trending field on the Y-Axis and Months will be the reference field on the X-Axis. If we want zone-wise analysis, we can add separate zones as Audit Units.

Click Forecasts and select 3 as we want a forecast for one quarter (i.e., three months).  We will get the expected marketing cost for the next three months following the pattern of the trend – provided marketing conditions and policies remain the same.