Challenge #103: Just another game?

[alacoume]

My solution:

[CeliaC_Silje]

Spoiler

[bkclaw113]

Spoiler

[olga_strubbe]

This was a fairly new exploration for me and the predictive tools I have are not a complete set.  I used @nick_ceneviva solution as a model.  Thank you, Nick, for helping me learn. 

Attached is my solution without the macros.  You can find macros in the standard solution supplied to this challenge, as I don't have all the pieces due to incomplete predictive tool set in my Alteryx Designer package. 

[DaisukeTsuchiya]

Spoiler

 

[StanleyKinnok_Chan]

It seems that forest model outperform the linear regression model. I come up with different conclusion because of this.

[geoff_zath]

Spoiler

[chandler-gjino]

attaching my solution

[SaiKrishna2589]

For the important predictors, another option is to use 'Stepwise' model to get the most important features .But I used Assosiate analysis and checking for multi-collinearity among predictors.

[ed_hayter]

Spoiler

Not sure I'd trust my model, although it suggests that Super Bowls see offensive under performance compared to expectation so that makes a bit of sense (and Defense wins Chips). So maybe it is good, but i would not use it to put any money on a SB.