Empowering Analysts with Self-Service Tools: Alteryx Breaks the Mold
There's no shortage of players that market self-service data prep tools. Only Alteryx is also in the "Visionaries" quadrant of Gartner's "Magic Quadrant for Advanced Analytics Platforms."
- By Steve Swoyer
- April 22, 2016
There's no shortage of players that market self-service data preparation products.
Only one of those players is also in the "Visionaries" quadrant of Gartner Inc.'s "Magic Quadrant for Advanced Analytics Platforms" report, however. That would be Alteryx Inc.
Recently, Alteryx announced a new "Predictive District" for its Alteryx Analytics Gallery. In Alteryx-speak, a "District" is a function-, application-, or vertical-specific collection of tools, workflows, and macros. Alteryx has Districts for Salesforce.com and Financial Services, among others. Alteryx's new Predictive District includes pre-built tools and workflows to perform A/B testing, affinity analysis, and survival analysis -- most of which have been built and submitted by users.
The tools and workflows available in Alteryx's new Predictive District augment a set of features and functions that have already received high praise from users and analysts such as Gartner.
Data Prep -- And More
Like rivals Datawatch Inc., Paxata Inc. and Trifacta Inc., Alteryx specializes in self-service data preparation software. Self-service data prep describes a kind of data engineering workbench aimed at both data scientists and business analysts, as well as other, non-IT (or, more precisely, non-data management-y) roles. "Data engineering" in this context isn't simply a function of cleansing, transforming, and joining data from multiple source data sets or systems. It has to do with designing data flows -- repeatable data engineering workflows -- that can be used as grist for analyses.
Alteryx is popular among users of self-service data discovery tools such as Sense (from Qlik Inc.), Spotfire (from TIBCO), and, especially, Tableau Software. Would-be discoverers use Alteryx and competitive products to prepare data so that it can be consumed in Sense, Spotfire, or Tableau. However, Alteryx also includes an advanced analytic platform for performing that analysis.
Regarding the people performing those analyses, Alteryx says it wants to focus on business analysts -- marketing analysts, customer insight analysts, and the like -- as much as data scientists. The popular caricature of the data scientist is that she's more at home with a command-line interface (CLI) or a text editor than with a GUI. Some say that she prefers coding her data transformations -- again, using CLIs and text editors -- in languages such as R, Python, and, in some cases, Java.
This is a caricature with some admitted basis in reality. However, as demand for data scientists continues to ramp up, more and more non-traditional candidates enter the field, and the available tools mature and become more self-service, this will change.
The point is that data scientists have used a variety of (often intimidating) tools to acquire and engineer data for analysis. Don't forget that an early application for Hadoop MapReduce was as a cheap and efficient platform for massively parallel ETL, i.e., data prep, on very large data sets.
Business analysts, on the other hand, have tended to be more GUI-dependent. The most powerful and popular data prep tool -- or data integration tool, for that matter -- is the venerable Excel spreadsheet. Excel is something the business analyst understands intimately -- a technology in which she's a virtuoso. The problems with using Excel for data prep are that it's far from ideal for many kinds of data manipulation/engineering and it can't be made to scale easily, elegantly, or manageably.
Alteryx aims to offer a user experience that's optimized for data prep and analysis and, moreover, scales to support reuse and process instantiation.
When you design a data flow in Alteryx, you're designing a repeatable process that can be packaged up, productized, and reused by other consumers. "Those people who are doing the traditional data science things, coding and R, they need to be able to scale what they're doing further, so rather than having to be the only person who's utilizing the insight that they've built, we allow them to take that content and embed it within a workflow," said Paul Ross, vice president of international marketing, during Alteryx's most recent presentation to the Boulder BI Brain Trust (BBBT).
On the one hand, Alteryx makes data scientists more productive, Ross explained. On the other hand, he continued, it's also designed for the more pragmatic needs of the business analyst.
"We're definitely not looking to provide a data scientist-specific tool," he told BBBT attendees.
"Our goal is to take the analyst, whom we believe has been incredibly under-served … and [empower] them, giving them some of those skills. We're definitely not looking to create a Python environment, by any stretch of the imagination. That's not our goal. Our goal is to take things like R and serve it up to an analyst so they can use it without having to have a PhD."
Business analysts probably aren't inclined to take up Python or Java, but they have been known to dabble in SQL. More enterprising business analysts -- which is another way of saying more frustrated business analysts -- have also deployed their own data mart and ETL technologies, too.
In other words, a lot of what a tool such as Alteryx does has in the past been performed by hand-coding SQL statements or by using commodity ETL technology, such as Microsoft's SQL Server Integration Services (SSIS). Almost all of us can point to cases in which frustrated analysts coded their own data transformations in SQL or acquired, configured, and used something similar to SSIS, but enterprise of this kind has always been more of an exception than a rule. For most business analysts, the alternative to self-service data prep is only Excel.
Alteryx even impresses analysts who are at home in SQL. BI developer Daniel Kresiva waxed enthusiastically about his first use of Alteryx on his Sculpting Data blog last year. Kresiva highlighted automated features such as intelligent join statements and multi-row/multi-field formulas, among others.
"Anyone who's worked in SQL for a while has made the mistake of trying to union two sets of data together that either had different numbers of columns, or columns that weren't in exactly the same order. The result is a bunch of errors, or even worse -- incorrect data," Kresiva wrote. "In Alteryx, unioning data is super easy. It doesn't matter what columns you have or what order they're in. You can just let it automatically configure the union based on field names or position. This greatly improves productivity and accuracy when mashing a bunch of disparate sets of data together."
Tableau Software Inc., Qlik Inc., and other vendors are beginning to catch up to Alteryx in some ways. At its Tableau Customer Conference (TCC) last year, for example, Tableau unveiled a similar "Union" capability. Also at TCC, Tableau announced support for cross-database joins. The latter isn't so much a data prep task as a function of federated query. At the same time, however, it mitigates -- or eliminates -- the need to go outside the Tableau environment to acquire and prepare data.
In spite of this, Tableau, Sense, Spotfire, and others, are likely to maintain a mutually beneficial relationship with Alteryx for some time to come. Moreover, Alteryx has its own base of -- by many accounts -- quite highly satisfied customers.
"Alteryx received very high reference scores for ease of use for citizen data scientists and business analysts in a code-free environment," wrote analysts Jim Hare, Gareth Herschel, Lisa Kart, and Alexander Linden in Gartner's "Magic Quadrant for Advanced Analytics Platforms" report.
"Alteryx complements the growing data discovery market by offering both data preparation and advanced analytics," they continued. "With ease of deployment to the Alteryx Gallery and a clear product and pricing strategy, customers see Alteryx as delivering business value. Customer satisfaction has risen from its already high level in 2015."
Stephen Swoyer is a technology writer with 20 years of experience. His writing has focused on business intelligence, data warehousing, and analytics for almost 15 years. Swoyer has an abiding interest in tech, but he’s particularly intrigued by the thorny people and process problems technology vendors never, ever want to talk about. You can contact him at firstname.lastname@example.org.