Alteryx Community Ideas

Atabarezz · ‎03-05-2016

Kaggle competitions are widely known in academia but not so much in the industry, at least in MENA region...

Mostly college grads, ML practicioners and analytics consultants are attending who frequently code... Instead Community can provide

kaggle tutorials completed on Altery,students and new users can get a grasp of advanced analytics easily...
solutions to old and recent challenges to be solved solely on Alteryx workflows...

To name a few tutorials;

Some recent interesting competitions are;

It may even be a nicer idea if someone publishes a top %1 solution in competitions using Alteryx,

discounts to inspire events or personal licenses may be provided or automatically becoming an ACE for 6 months/a year etc.

JohnJPS · ‎03-05-2016

I would be very interested in this. Right now an Alteryx-only solution would be hard to place highly. However, if we could gradually build macros to bring Alteryx along to where it can do automatic parameter tuning and/or model ensembles, that would be pretty exciting. Feature analysis/reduction/engineering are huge too in Kaggle, so whatever general tools/macros we could build around those concepts would help too.

(But honestly for me, time is the biggest enemy for Kaggle.)

Atabarezz · ‎03-05-2016

Random forests and xgboost seem to be the most common methodsthat bring top places in kaggle competitions...

Of course there are many excentric algorithms;

jacknife regression, deep learning,
dbscan, autoencoders, self-organising maps

Though I guess it's not that hard to implement most fresh R packages...

Probably the best part will be challenging Alteryx community on cleansing kaggle data,

enriching it with weather, spatial or social data...

Best

TaraM · ‎04-11-2016

Very interesting @Atabarezz. I'd love to hear from our community users if there is interest and what their ideas are around Kaggle competitions and datasets and using Alteryx.

Atabarezz · ‎05-07-2016

A quick stat from the community;

Kaggle has been refferred my multiple users, so there is growing interest on using Alteryx for kaggle competirions;

Ideas by @MarshallG and @JohnJPS referring to the succeeding xgboost method in kaggle competition

http://community.alteryx.com/t5/Alteryx-Product-Ideas/Xgboost-Regression-as-a-predictive-macro/idi-p...

Here is @kgolynko and @Riotsolving the san fransisco problem on kaggle using Alteryx;

And here is a an effort to replicate Titanic model by @vaibhav_jain

http://community.alteryx.com/t5/Advanced-Analytics/DT-Titanic-Dataset-using-R-pallet-in-Alteryx/m-p/...

I sincerely believe we need a kaggle/crowdalytics community category as well as a specific district that contains data sets and examplary Alteryx workflows to get going...

What do you think... Start that pls if you like...

Best

Cheers from İstanbul

JohnJPS · ‎05-09-2016

If anyone's interested, the workflow attached in my post here, actually generates submissions for the Kaggle Titanic challenge, using both GLM and GBM approaches with the Alteryx predictive tool versions thereof. It also imputes some missing values and excludes some uninteresting columns (based on field importance observations from the GBM tool). These are just baby steps in the Kaggle world, but decent first steps nonetheless.

MarshallG · ‎05-10-2016

While first learning Alteryx (and playing with some of the models like Random Forest for the first time), I certainly used the Titanic survival dataset and resources from Kaggle. Like @JohnJPS suggests, time is the limiter for me in terms of submitting to competitions. I think that Alteryx would certainly be helpful in the data cleansing and feature engineering steps. To actually be competitive at the model building steps, I think we would likely have to be running R scripts from Alteryx -- which is fine -- but certainly requires a higher level of programming than many community membmers possess (I include myself here. I can hack away to create a passable linear or logistic model but have never tried to create an XGboost model, which seems to be a part of every successful ensemble model at this point). All that said, I'd be happy to have a forum/gallery for Kaggle competitors using Alteryx for when I have some spare time at work ;-).

I also think that if someone could get near the top of some leaderboards exclusively using Alteryx, it might generate some marketing buzz for Alteryx-- in particular with younger, aspirational data scientists who are currently using open source tools.

andrewdatakim · ‎05-10-2017

I just joined the Kaggle community and I can't wait until a Python integration has been completed with Alteryx between Python, R and Alteryx I feel I make a run for some of the competitions. @JohnJPS or @Atabarezz do you know if we could DataRobot for Kaggle?

JohnJPS · ‎05-10-2017

@andrewdatakim,

My understanding from afar is that DataRobot could certainly be used for Kaggle. I don't have a license for DR though, so haven't had the opportunity to play around with it.

Cheers,

John

Atabarezz · ‎06-28-2017

https://techcrunch.com/2017/03/07/google-is-acquiring-data-science-community-kaggle/

This is a news from previous quarter... But it proves the value in creating an Alteryx version...

1. corporations and medium sized businesses can put their business problems to test sharing their anonymized data...

Data can be prepared and

anonymized in Alteryx as well and most probably will be...

2. Academician, Alteryx artisans, consultants, partners will try their best to solve and share their "Alteryx" solutions

Lot's of alternate solutions

Nice best practices

New business solutions...

Best

please support the idea, share it with friends and colleagues...

WillM · ‎08-09-2021

Alteryx Community Ideas

A community category for Kaggle/Crowdalytics