Thank you to all who attended our Alteryx NYC User Group Meeting on Wednesday Feb. 8th @KPMG in Midtown!
I enjoyed speaking with many of you and feel we were lucky to have 2 excellent presentations.
First, Suresh Vadakath, Senior Solutions Engineer @Alteryx reviewed the date tool (always a popular topic for those new to Alteryx), as well as utilizing the output tool to parse data into multiple files and send out via email. (I have personally used this in a pinch to send custom reports to 36 district managers, huge time saver.)
Suresh then gave us a great overview of Alteryx's tight integration with DataRobot, an automated machine learning platform. https://www.datarobot.com/
Both during the presentation and during side discussions, the topic of Alteryx and machine learning was raised.
Depending on where your organization lies on the data science spectrum, Alteryx can help in multiple ways.
The first way in in which Alteryx can assist in predictive modeling is in the laborious task of data preparation and feature engineering.
It is quite common to offload this work to Alteryx and then feed the prepared data sets to Alteryx's own predictive tools (based on R) or to pass this output to models built outside or Alteryx in R or Python or even now to an automated ML platform such as DataRobot.
Using Alteryx for the preparation step has 2 major advantages. It is fast + it is shareable across more teams/personnel.
Business Analysts can do similar preparation/imputation that data scientists typically do with packages such as dplyr or pandas.
This can free up your data scientists to do more actual modeling which benefits everyone.
A nice overview of the usage of Alteryx with DataRobot synergy is covered in a webcast by Mark Haydoutov of S&P Global.
https://www.datarobot.com/alteryx-datarobot-webinar-recording/
The whole thing is a good watch but the 32nd minute has some of the best content.
Our second presentation was by Arthur Franke - Manager - Data & Analytics @KPMG and a recovering particle physicist;>
Arthur presented his awesome Data Validation Toolkit. Arthur and his team used a combination of Alteryx workflows to tackle a common problem.
Given a brand new data set, how can I learn as much about the data as possible, as quickly as possible? The Toolkits systematic approach details tons of info about the data (primary keys, data distribution, cardinality, identify outliers, investigate data quality, the ability to pivot on anything etc.) Arthur mentioned how this tremendous time saver had features so popular that internal teams were requesting to download the features for actual work during the beta testing. The kit comes complete with version control and color coded documentation and is 1 of the most complete data process/solutions I have seen built with Alteryx.
All in all, this was a fun meeting.
Special thanks to Katherine Palmer & KPMG for once again hosting.
Aaron Walters (CME Group), Katherine and I all look forward to our next meetup in April.
Look out for user group communications where we will be looking for your feedback on the topics you want covered plus welcome anyone who wants to present their workflows.
Best,
Dennis Peters
Alteryx NYC User Group
https://community.alteryx.com/t5/New-York-City-NY/gp-p/nyc-metro
