Interview with Alteryx's Chief Data Scientist Dr Dan Putler

by Angela Ashton

Alteryx's Chief Data Scientist shares his thoughts on the future of predictive analytics while providing practical insights on how to get started with Alteryx. 

1.What does the future of Predictive Analytics hold for Alteryx? Alteryx has seen fantastic growth over the past few years, what do you see as some of the key business drivers within your customer base that have fuelled this growth?
 
I think one of the key factors is the realisation within the customer base of just how well Alteryx works and how sticky it is with the user base once adopted. Alteryx has always had an exceptionally good reputation within its user community particularly within two key segments:
 
a) Those who have been trying to do far too much with excel spreadsheets for far too long and who are now seeing that house of cards coming down on them. There is only so much you can do with pivot tables, VLOOKUPS, and linked spreadsheets before it all becomes a horrible mess and you are unable to go any further beyond a certain level of complexity. Alteryx not only replaces these excel based processes but provides additional capabilities that allow users to get into spatial and predictive analytics. It also offers a significantly easier interaction with back-end systems which means the resulting processes are easier to document and follow. What all this essentially means is that Alteryx is much easier and a lot less cumbersome to use than their previous processes, which is what drives user adoption and expansion. 
 
b) Those using Alteryx to replace legacy software which tends to have high license prices associated with them. For instance, I have found many customers replacing SAS processes with Alteryx based ones because they are more robust and better able to connect into back-end. The breadth of things that can be done with Alteryx is enormous - from basic data prep right through to spatial and predictive, to prescriptive analytics - and because it follows the same basic paradigm it is comparatively easy to use with a significantly less steep learning curve than legacy systems. The code-free and code-friendly aspect of the product are also what make it an ideal solution for many organisations, as we can address the needs of both emerging analytic talent who is exploring more sophisticated analytic processes without requiring heavy SQL, R or Python coding knowledge, whilst also supporting the statisticians and trained data scientists who need robust data processing and modeling functions and prefer to code. Because we’re able to satisfy both user needs, our platform bridges the analytics language gap that often divides these two teams, bringing them together and empowering them to collaborate in ways that other platforms just can’t.
 
 2. What are some of the key business drivers that most excites you about the future growth of Alteryx?
 One of the biggest business drivers we’re seeing today is the hyper focus on digital transformation, which is extremely exciting as analytics and analytic teams serve as the backbone in supporting transformational initiatives. This means faster, more agile development of insights, across teams, users and departments. It also means organisations are looking to a modern approach to analytics, one that isn’t made up of a stack of disparate tools, and processes that require different skills and are built for different users. The organisations who use Alteryx are directly benefiting from the simple, agile and modern approach to analytics. They recognize how easy we slipstream into their existing analytics ecosystem and unify the analytics process. As a result, we continue to expand our functionality to further reduce the number of software tools that knowledge workers need to rely on so that analysts and data scientists can work increasingly exclusively within an Alteryx based context which will ultimately make life much easier for them. And more importantly help shorten the distance between business question and analytic backed insights. What is exciting is that these drivers are not going away - the past drivers are the driving elements of our future.
 
 3. What is on the Alteryx roadmap? Are you able to share some key highlights of the Alteryx roadmap with us?
 Alteryx Connect - the functionality within Alteryx Connect will become both broader and smarter as we move forward, particularly with respect to the socialisation of data discovery so that people better understand what data assets are available to them. We believe that tribal knowledge and the sharing of work that analytics teams are creating is going to be critical in this next wave of analytic evolution. The more pervasive analytics become within an organisation, the more critical it is going to be to share information across analytics teams as a means of preventing recreating the insights wheel every time a business question is asked.
 
 From a predictive analytics perspective there are two key users that we focus on:
  1. The Line of Business users that would like to get into predictive analytics but find it somewhat daunting. These are the users that we have been thinking about for a long time, and they are very front and centre regarding what we want to focus on in the future.
  1. The more experienced users who have been using predictive analytics for some time and who generally fall into the following two camps - those that are coders and those that are not.
 From this perspective, our aim moving forward is to ensure that we are setting our products up to be both code free and code friendly. What we are working on going forward is that in addition to having R functionality within a direct workflow coding base while also allowing for Python. This is similar to the announcement we made some time ago in the context of Spark in that we have the apache spark tool which allows users to write code directly in either Scala, Python or R. We are also paying a great deal of attention to the tools that coders are currently using and how we can take advantage of those tools in the Alteryx context.
 
 4. Why do you feel that users are often challenged by making the leap from BI to Predictive Analytics and what is the best way to deal with it from a software perspective?
 We often find that it is the mind set that needs to be addressed with those that are making the transition from traditional BI environments and in particular the awareness of just how easy Alteryx is to use. The Alteryx way of doing things is very transparent and intuitive which means users experience huge productivity gains with a minimal learning curve.
 
 Getting Started with Alteryx
 
5. We often see customers who are reluctant to embark on a predictive tools path because they feel they are not large or sophisticated enough to manage this. What advice would you give these sorts of companies to help them make a start with Alteryx?
 The best advice I can give to any organisation regardless of size or complexity is to adopt an organisation structure (based on a set of best practices) that will allow them to support and nurture an in-house data analytics capability. What makes the adoption of a data analytics easier, particularly for smaller organisations is to choose a set of tools that offer a comparatively less steep learning curve. Some of the most successful data analytics stories have been born in relatively small companies who have adopted a strong data analytics culture and supported it via the appropriate business structures.
 
 6Why shouldn’t companies be reliant on R alone?
This goes back to the code free v's code intensive question. For organisations who need a large number of their users to perform predictive analytics, R is going to be a tough choice as it requires knowledge of a different software platform to the one that users would normally use on a day to day basis. Additionally, Alteryx offers huge speed advantages over R regarding getting data ready and optimised for use, which means that power users can spend significantly more time on data analysis v's the time they would otherwise have spent on data prep.
 
 Alteryx in the Office of Finance
 
7.Are there any opportunities for machine learning within the Office of Finance?  If so can you walk through some use cases for us?
 When you think about the Finance department, it underpins every other business function across an organisation, so the applications for Alteryx in that scenario are almost limitless from inventory planning, cashflow planning, labour costing, sales planning, revenue projection planning right through to attrition management and retirement planning.
 
 About Dr Dan Putler
Chief Data Scientist at Alteryx, Dr Dan Putler has over 30 years of experience in developing predictive analytics models for companies and organizations that cover a large number of industry verticals, ranging from the performing arts to B2B financial services. He is a co-author of the book, “Customer and Business Analytics: Applied Data Mining for Business Decision Making Using R”, which is published by Chapman and Hall/CRC Press. He is one of the principal shapers of Alteryx's roadmap for predictive analytics and machine learning, helping to guide Alteryx into the Leader quadrant of Gartner's Magic Quadrant for Data Science and Machine Learning. Prior to joining Alteryx, Dan was a professor of marketing and marketing research at the University of British Columbia’s Sauder School of Business and Purdue University’s Krannert School of Management. 
 
 About Alteryx
 Alteryx is revolutionising business through data science and analytics, and we are empowering everyone in an organisation to experience the thrill of getting to the answer faster. Our end-to-end platform unifies the analytic experience, enabling organisations to break data barriers. The Alteryx Platform provides the analytic flexibility that business analysts, data scientists and IT need to discover, prep, analyse, and operationalise analytic models through a collaborative and governed platform.
 

Recent Articles

Blog Categories

See all

Search