Challenge #229: Trend Analysis

[tristank]

Hey this was pretty fun! I had no idea how to calculate trend and am terrible at following instructions so I had Chat GPT walk me through it step by step.

Spoiler

[gawa]

Done

[ed_hayter]

Spoiler

Had parked this challenge as i first looked at it on a friday afternoon and didnt fancy the maths. In the end it was a really fun and engaging challenge.

[alacoume]

Trend is just linear regression!

Spoiler

 

[Hub119]

Not the prettiest of output "reports" but it gets the job done in a pinch.  Solution attached.

[asultanov]

Spoiler

[olga_strubbe]

The formulas were quite challenging for me on coming up with the trendline; I, therefore, leveraged the logic on the trend calculations from the solution provided. Thank you, @AYXAcademy, for showing how it is done!

Please see my solution attached.  

[Rob-Silk]

Solution attached below:

Spoiler

[Erin]

I am starting my predictive tools journey and knew this one could be accomplished with them, but had to lean on some posted solutions for it to click. Very thankful for @paulfound's post and both of @mustel's posts!

Three routes w/ three Predictive Tools:

Spoiler

First route uses the Pearson Correlation Tool (technically on the Data Investigation pallet but shouldn't do Predictive without Data Investigation first, so they're related) to find r, and a summary tool to find the standard deviations and means. Then a formula tool uses these variables to calculate a and b in y = a +bx as well as apply this formula to the data points. Boom - trendline. Makes you really appreciate models that calc this automatically! 

Second route uses the Linear Regression Model report output to find a and b, where a is the Estimate intercept and b is the [predictor variable] Estimate. This route runs the models once and then manually plugs them in the y = a + bx formula. 

Third route is a fancier more dynamic version of the second route. It adds the Score Tool to calculate the trend line. 

[Bobbyt23]

Heavy maths for a Sunday afternoon.😜