Challenge #127: US Grand Prix Lap 2 - Employee Retention

We're on to Lap 2! At this point, our drivers have finished Lap 1 and had a chance to shake off the nerves and prepare for the next lap.

The first lap had our drivers prep and profile a data set that we will use in Lap 2. This lap features the predictive tools and their use for aiding in employee retention.

Like with Lap 1, if you would like to follow the format of the Grand Prix, give yourself 2 minutes to read the prompt without touching the mouse or keyboard. Our drivers had 12 minutes to complete this exercise. Good luck!

Pictured below are our finalists @BenMoss, @mattagee, and @NicoleJohnson trying on the Grand Prix helmet.

challenge_127_start_file.yxmd

Preparation

Intermediate

Predictive

Accepted answers

JimmyG

Here is the answer I showed at Inspire. Good job everyone!

The final lap posts next week - get your spatial skills ready!

challenge_127_end_file.yxmd

All comments

ed_hayter

FYI - think the forest model returns different results depending on Versions. I was very confused why my set-up was not returning expected outcome and read through some comments from earlier that suggest the same thing

Spoiler

challenge_127_start_file.yxzp

geoff_zath

I got close running 23.1. It looks like the Decision Tree might have changed since this challenge was originally made.

Spoiler

challenge_127_GKZ.yxmd

gawa

done

challenge_127_start_file_gawa.yxmd

bo_peng2022

my solution

challenge_127_start_file_solution.yxmd

daiphuongngo

Spoiler

Challenge #127 US Grand Prix Lap 2 - Employee Retention.yxmd

IgnacioVilaplana

Attached my solution! I had to change the filter of employees with more than 50% probability of leaving since with the 23.1 version I'm using the best model doesn't think anyone is above that threshold, so I lowered it to 20% to get the riskiest employees.

Spoiler

challenge_127_solution_IVS.yxmd

Rob-Silk

Solution attached below - as many others have already pointed out, the Decision Tree model still gives the best accuracy, however, the results the tool produces no longer seem to be the same as the original output in newer versions of Alteryx.

Spoiler

challenge_127_solution_RS.yxmd

olga_strubbe

Thank you, @JimmyG! This was a good practice for me using Predictive tools.

Challenge 127_US Grand Prix Lap 2_Employee Retention.yxmd

Erin

Wow I was racking my brain trying to figure out where I was going wrong, and turns out the models have changed over the years with the new versions. Phew! Would be cool if this could be reworked somehow so that we can work it with newer versions of Alteryx.

challenge_127_submission_EM.yxzp

lynnesonney

Curious if I'm getting a different answer just because of versioning and R updates or if I'm missing something, I swear my config is the same as some of these other solutions which seemed to work. I even tried switching on the oversampling the in the score tool - no dice. Please let me know if you see something I don't!

I actually downloaded some solutions and ran them as is - now they don't match the solution! going to to chuck it up to a versioning issue. I'm on 2023.1 - maybe something to consider when going back 6 years to find practice problems....

challenge_127_lsonney.yxmd

Dibas98

Solution

challenge_127_start_file.yxmd

cusher

Spoiler

Versioning of the model tools causes the original flow asnwer not to match the results from the model tools.

JoachimCaronTIL

Here is my solution.

I did not get the same results as yours as the model have changed since this challenge has been published...

Spoiler

challenge_127_JC.yxmd

Towers

So it seems the models have changed and its not possible to get the same answer as the original challenge 😭, Id love to had known that before adding an extra 3 minutes onto my time 😅 fun one regardless

Spoiler

challenge_127_towers.yxzp

BS_THE_ANALYST

Was frustrating as the answer doesn't match.

The irony here is funny. Predictive challenge, and none of those people match to the answer (at least for me).

Spoiler

challenge_127_start_file.yxmd

SJKIM

Done

challenge_127_start_file_Done.yxmd

lwolfie

Same as others, it would be nice if the answer were updated. I don't use Predictive tools, so outside of copying others solutions, it's hard to verify my solution.

challenge_127_solution_LW.yxmd

Bobbyt23

Good to practise those tools but couldn't get the final answers as others have said.

Weekly Challenge 127 18-10-24.yxzp

cmoussa

challenge_127 CTM.yxmd

Carolyn

Same as others, couldn't get the answer to match. What's interesting is that if I include Days_of_Employment, I get no answers in the Filter. If I exclude it, I get some answers. They don't match the solution, but it's at least not zero

Spoiler

challenge_127_Carolyn.yxmd

JBevan89

Like everyone else, the answer is different now to a different dataset

challenge_127_JB.yxmd

Quick Links

This months top contributors

mceleavey 383

mbarone 337

Hollingsworth 335

LanisC 335

JeffF 335