Tableau has a built-in Forecasting feature that attempts to predict future values based on data patterns.
See it in Action:
One should get an idea of how a calculation is formatted by software. From Tableau’s site, we get insight into how they predict future values:
Forecasting in Tableau uses a technique known as exponential smoothing. Forecast algorithms try to find a regular pattern in measures that can be continued into the future.
You typically add a forecast to a view that contains a date field and at least one measure. However, in the absence of a date, Tableau can create a forecast for a view that contains a dimension with integer values in addition to at least one measure. For details on forecasting using an integer dimension, see Forecasting When No Date is in the View. – Tableau’s Site
Tableau by default analyzes dates and looks for seasonal changes. Meaning, if sales by year pick up a few months before December and trail off in February, Tablaeu by default analyzes that seasonal pattern. This can be burned off in the Tableau Forecasting Options.
Tableau’s online documentation details how forecasting is computed using exponential smoothing:
The simplest model, Simple Exponential Smoothing, computes the next level or smoothed value from a weighted average of the last actual value and the last level value. The method is exponential because the value of each level is influenced by every preceding actual value to an exponentially decreasing degree—more recent values are given greater weight. – Tableau’s Online Documentation
These default settings can be modified in the Forecasting Options.
Taking the L.A. City data on crime statistics (which I used in a prior visualization), I constructed a simple forecasting report.
The above figure (click for full size) shows predictive forecasting for a date range of a specific crime. Here in the graph we have the crime “Assault with a Deadly Weapon” over a date range of June 2010 – Aug. 2010.
As long as we don’t have missing data in the date range, forecasting does a pretty good job of future predictions.
There’s a faded line of the future date prediction, which has a blue boundary of potential variations of data (called “Prediction Intervals.”)
Under the Analytics tab is a sub-section called “Model.” Within the Model sub-section is the “Forecast” option.
Simply drag the Forecast option onto the graph. Once you drag it onto the figure, the Marks window will add “Forecast” to the color settings.
The contextual menu, as a sub-menu of its own which offers the option: “Forecast Options.”
If you want the statistical details of the Forecast you applied, you can gather this in the Forecast contextual menu and clicking the sub-menu “Describe Forecast.”
This will give you a window with two tabs:
Here’s an example of Summary Statistical Data:
Below is an example of the Model Statistical Data: