My question was answered by Tableau Sales Consultant.
The Trend line metadata is currently not available to the user. You can see the data when right-click on the line and selecting "Describe Trend Line", but results are not available.
My workaround will be to replicate the 'Value' via a Calculated Field and use it to determine EOY projections. Once completed, I will post and share.
1 of 1 people found this helpful
I was able to calculate the metadata needed by using a Linear Regression Analysis formula via SQL, prior to uploading Data in to Tableau.
Using various queries, I shaped my data to where X = 'Dates' and Y = 'Running Sum Total' then applied the formula below
Round((Avg([X]*[Y])-Avg([X])*Avg([Y]))/VarP([X]),6)) AS RegressionCoefficient
A stronger Tableau user could create a calculated field to accomplish the same task; I'm not there yet. Please let me know if someone figures it out, Thanks
I was just looking for the same answer, and found this Interworks blog article. I thought it might be helpful for you as well. It tells you how to do trend line analysis with Tableau and R.
Glad to see you were getting something to work. I have a question, that apparently is so simple, it is glassed over by all knowledge articles as they dive into the deep stats around it.
When you say you set X = 'dates', how are you representing that numerically? I have a trend model equation : ln(per min concurrency) = -13.2465*(Minute(CreateDate)) + 577823, but I'm not sure what values I can actually insert to represent (June 7th @ 12:30 PM). Do we use unixtime? Something else? How do I translate minutes into a generic number in such a way that I can use it in the formula?