What is Data Preparation and understanding Vendor Capabilities

I guessed I missed the memo saying that ETL, Data Integration, and Data Quality are outdated terms (and technologies) and the new term used at least by Gartner and at O'Reilly Strata last week in NYC is Data Preparation. I can't say that I blame the industry because few executives understand data integration, ETL is associated with linear, long running data batch processing, and last generation's enterprise versions of data quality tools were too complex and expensive for many organizations.

What is Data Preparation?

At Strata last week I was able to see a number of vendors that are listed in Gartner's Guide for Self-Service Data Preparation. What is universally different from the new generation of data tools versus older ones is that they are "self-service" tools targeting either business users, data scientists, or citizen data scientists. Beyond administrative tools, the capabilities and user experiences of these tools are largely designed for non-technologists, although many enable advanced capabilities and the ability to plug in R, Python, or other code developed by IT or data scientists. This is a sharp contrast to ETL and data quality tools of the past that were largely designed for IT. Gartner goes on to illustrate several other differences around data types, sources, models, and other areas of differences between traditional data integration and data preparation.

When you review different vendors, you'll see that they focus on different aspects of data preparation and the sophistication of their capabilities vary considerably.

  • Many market themselves as tools that enable Big Data and Hadoop by allowing business users to easy load and join data sets into their Hadoop clusters. For the CIO that has big data infrastructure that is underutilized, these tools are designed to be "on ramps" to enable business users to begin loading data and performing analytics. DataMeer, Pentaho, and MicroStrategy are examples.
  • Some are positioning themselves as data lakes, or data lake management tools. They have similar data preparation capabilities for business users, but add IT operational capabilities around the data lakes. Vendors include Zaloni.
  • Most vendors then have some capability to either transform, join, or cleanse data. Many will use visual pipeline tools to illustrate steps, data grids to profile and identify cleansing issues, visualizations to help users discover basic information around the data set, and catalogs so that users can search and find relevant data. Example products are from Trifacta, Talend, Informatica, and Paxata
  • Many of the data visualization vendors also have some amount of data preparation capabilities. For example, Qlik can do data transformations and Tableau can easily do blends, grouping and binning. Other vendors in this space are Microsoft Power BI and SiSense.

I didn't get to visit every vendor at Strata such at Attivio, Waterline, ClearStory Data, RapidMiner and others that play in this space, but it's clear that CIOs and Chief Data Officers have their work cut out for them reviewing and selecting data platforms that work easily with their existing investments and platforms.

Want to learn more? Please sign up for the newsletter where I will share one secret to evaluating data preparation platforms.

No comments:

Post a Comment

Comments on this blog are moderated and we do not accept comments that have links to other websites.