Timeseries Forecasting Using Compare Tables!

You want a quick easy way to predict future data points? This exercise will allow you to create a fancy linear extrapolation with 4 easy steps to follow....

For this exercise we will use our Opportunities dataset, group it by Close date (Year-month) and extrapolate a measure so we can forecast that into the future! Fancy right?

So let’s get to it!

STEP 1:

The first step is to select your Opportunities dataset and open it in a new lens and do a Count of rows to get the total number of Opportunities and group your Data by Close Date (Year-Month)

New Lens > Opportunities dataset > count of rows to get total no. of Opportunities grouped by Close Date (Year-Month)

STEP 2

Change your lens to a compare table and go to Formula > f(x) > select timeseries from the dropdown > select a measure from Column: which in our case if Count of Rows > specify 6 prediction points > seasonality: Auto

Compare table > f(x) > Timeseries

STEP 3

Once you have created your timeseries, you will add a Coalesce- which will allow us to do a Combination of Actuals and Predicted.

A coalesce COALESCE (A,B)- Specifies the first value if it exists (first non null value), and if so displays it. If not, takes the second one (predicted one)

STEP 4

Hit apply and you will format your chart-

1. Hide the original measures- including your original count and your prediction

2. Display the coalesced value
   Then prolly switch to a timeline as opposed to a combo - that'll guarantee you have the predictive line option in your formatting menu

3. Finally enable your Predictive line and change the color or formatting

New lens displaying timeseries

Formatting to enable the Predictive Line on your chart

Final Chart with the Predictive line to display where are you now vs the future predictions