Can you help me beat people using Python to make predictions?

[Gualigee]

Hello, 

I participated firm's challenge. I currently got an F1 score of 0.74. The highest is 0.78.

If I remove the yellow highlighted below, my score dropped to 0.73.  

If you have time, do you mind looking at my workflow?

https://uofi.box.com/s/fb36fm8kgo8p3cwocsku4begrz08xqco

Any advice would be greatly appreciated. 

Trigger warning: big data is coming up!

I believe I am the only one who uses Alteryx.  

Can you help me beat people using Python to make predictions?

Thank you very much for your help. 

[TheOC]

hi @Gualigee 

Please can you attach your workflow to the post, rather than the link to uofi.box? I have not seen that link before, and do not click links i do not recognise. 

Would be happy to help you with your predictions. Remember that Python can run inside Alteryx, so Alteryx will always win or be as good as ;) 

Thanks,
TheOC

[MarqueeCrew]

@Gualigee ,

What you'll want to do is:

OPTIONS: 

>> Export Workflow

This will create a zipped file (yxzp) that contains all of you INPUT and OUTPUT data along with any macros that you've created.  Then you can attached that file to the post.

Cheers,

Mark

[MarqueeCrew]

P.S. What does @TheOC get if he beats the .78 score?

[TheOC]

@MarqueeCrew Oooo is there a prize? I love prizes... 😁


Cheers,
TheOC

[MarqueeCrew]

@Gualigee ,

By the way, I like to reduce redundancy in my IF statements.

If you are older than 100 then "You're really old"

Else if you are 100 or less and you are bald then "You've lost more than your youth"

Why would you worry about the person being less than 100 in the second IF statement?  You already know that they are 100 or less.  That's a given.

IF [occurrenceStatus] > .948 THEN 1
ELSEIF [decimalLatitude] > -30 THEN 1
ELSEIF [decimalLongitude] > 147 THEN 1
ELSE 0
ENDIF

But really, this is an OR condition with:

IF [occurrenceStatus] > .948 OR [decimalLatitude] > -30 OR [decimalLongitude] > 147 THEN 1
ELSE 0
ENDIF

Cheers,

Mark

[Gualigee]

Hi @TheOC , thank you for your reply. 

The source data is huge and exceeded the limitation of the post. 

The box link is safe. You can click on the download button to get the data and workflow. 

Thank you. 

[Gualigee]

Hi @MarqueeCrew , thank you for your reply. 

The source data is huge and exceeded the limitation of the post. 

The box link is safe. You can click on the download button to get the data and workflow. 

Thank you. 

[Gualigee]

Hi @MarqueeCrew 

IF [occurrenceStatus] > .948 OR [decimalLatitude] > -30 OR [decimalLongitude] > 147 THEN 1
ELSE 0

I agree with you. Thank you for the advice. However, not every part is ">".

if [ocurrenceStatus_1]>.96
then 1
elseif ([ocurrenceStatus_1]<=.96 AND [decimalLatitude]>-30 OR [decimalLongitude]>147)
then 1
else 0
endif

[Gualigee]

haha, i will give you a prize.