Forecasting is a process of predicting the future value of the measure. We are going to use regular patterns in the measure value to predict future trends.
The technique of identifying regular tends from the existing data value and adding a forecast is known as Expontanatial smoothing.
There are multiple numbers of exponential smoothing models, in that, Tableau selects the best quality model as a final forecast model.
The forecasting is based on two important things such as Trends and Seasonality. The Trend is an increase or decrease in data over time. Seasonality is the repeating variations in values over a determined period of time like Years, Monthly, Quarterly known as seasons.
Tableau gives us a number of options to select from, for a forecasting model. Either, we can keep it on Automatic and let Tableau decide which forecasting model is the most appropriate for our data and requirements.
Or, You can choose the Custom option and select from the three types of trends and season characteristics which are listed below:
- None : This option selects none of the model components.
- Additive : The additive is the one in which contributions of the model components are summed.
- Multiplicative : The multiplicative model is one in which at least some component contributions are multiplied. Multiplicative models can significantly improve forecast quality for data where the trend or seasonality is affected by the level (magnitude) of the data.
Filters in Tableau
Follow the below steps to create a forecast in Tableau:
Make sure you are using at least one date dimension and one measure variable in order to create a forecast.
- Open Tableau and click on Connect to Data.
- Then the Connect page will open, click on Microsoft Excel.
- Navigate to the downloaded files, select Sample Superstore.xls excel file, which is a saved excel file, and click on Open.
- You can download a Sample Superstore.xls file from the link: https://chercher.tech/files/
- Open Tableau and connect Sample Superstore data source.
- Once Tableau has connected, click on Sheet1, to navigate to the worksheet.
- Create a simple line chart by dragging Order Date to the column shelf and Sales to the row shelf as below.
- Click on the Analytics pane and select Forecast under Model.
- Drag the Forecast option into the graph area to apply it on our line chart.
- Click on the + sign in Year(Order Date) to convert it to Quarter(Order Date), you can see that the graph lines has been changed to quarterly, and Forecast will immediately add an extension line to your line chart as below.
- Drag and drop Sales to the Label shelf, you can able to see the predicted values.
- In the below graph, you can see the actual sales rate.
- And, in the below image, you can see the estimated sales value by the year 2018 4th quarter.
- You can explore more options related to forecasting by right-clicking on the forecasting section and go to the Forecast option and select Forecast options.. from the list that opens further.
- Once you select forecast options, the below dialog box will open, in this dialog box, you can select and set the forecast length, source data, aggregation type, and the forecast model type.
- The forecast model is usually set to Automatic as it gives the best and the most appropriate forecast graph. You can also select or deselect the Show prediction intervals option and adjust it as per your liking.
- You can also view a detailed description of the forecast by right-clicking on the forecast section and then go to Forecast and select the option Describe Forecast.
- The below image shows the details of the forecast Summary.
- In the below image, you can the Models of Forecasting.
- You can change the color of the forecasting area, click on the color card on the Mark shelf, and select edit colors options.
- Now Edit Colors window will open, click on the Estimate and select the color from the color palette and click on Assign palette and then click Ok.
- Now, you can see that the color of the forecast line has been changed.