- Bernard K
How to perform forecasting in Tableau
Forecasting is the process of making predictions of the future based on past and present data and most commonly by analysis of trends. Forecasting in Tableau uses a technique known as smoothing. Forecast algorithms try to find a regular pattern in the measures that can be continued into the future.
Requirement for forecasting
To create a forecast your view must be using at least one date dimension and one measure field. However, when a date field is not present in the view. Tableau looks for a dimension in the view which has integer values. When Tableau finds such a dimension, it uses it to predict additional values for measures in the view.
Note: Forecasting cannot be added when the view contains table calculations, disaggregated measures, grand totals and sub-totals, date values with aggregation set to exact date.
Creating a forecast
In the example, I am using Superstores data set to forecast Sales in the next 12 months. To do so, I will be using Order Date at Month-Year level of detail and Sales.
First created a basic line chart.
Add forecast using one of the following steps:
Under Analytics pane, drag Forecast to the view or
Under Analysis menu >> Forecast >> Show Forecast
Tableau returns forecast indicator – showing the actual data extended by forecasted data.
We use forecast options dialogue box to configure forecast options such as:
Length of forecast
The range and temporal aggregation of the data source from which to generate the forecast
The forecast model
To open forecast options dialogue box >>Analysis menu >> Forecast >> Forecast Options….
On the forecast options dialogue box, you can customize your forecast.
(In this case have chosen the length of my forecast to be exactly 12 months)
Creating effective forecast is necessary for sound planning. Giving confidence to managers when making important decisions.
Stay tuned for more exciting content on Tableau. To receive more of the Tableau tips and tricks, kindly join our mailing list by subscribing below.
Thank you for reading.