Art of Analytics, Chabot and Stolte Keynote with Devs

“At Tableau, we believe that empowering people to unleash their creative genius is the most important goal which drives modern analytics strategy.” So says Tableau CEO Christian Chabot in his keynote to open the 2014 Tableau Conference to an audience of data analysts, engineers, investors and journalists.

Chabot underscores this describing data analysis as a creative process, and that it takes conditioning for people to become creative thinkers in modern society, but, more importantly for the attendees at a data analytics conference, creative problem solvers.

He builds to this point in kicking off the seventh annual edition of the conference – previously the Tableau Customer Conference – by noting the growth over the event’s history: 180 the first year, then 300, 70, 1200, 2000, to last year’s 3,000 paid attendees, and over 5,500 this year. Although they may come from places as diverse as London, New York, Tokyo, and Miami, they are a community, connected by their passions for data.

Several members of this community are called out for their passion which extends beyond the workplace. Among them:

• Matt Francis, who works in genomics by day and charts astronomic data – like how the sun influences weather patterns – in his time away from the office.
• Ramon Martinez, of the Pan American Health Organization, who blogs about global health issues.
• Kelly Martin, a self-professed “grandmother data geek” and former history teacher who publishes data on oil spills.

Not all the passions are as serious as the environment and disease pandemics. Ryan Robataille has published data from an official database of US Bigfoot sightings. Apparently, there have been 424 “Class A” sightings around the Seattle metropolitan area, so hikers are advised to exercise due caution.

Chabot pivots from the community, and addresses the mission of Tableau and how that dovetails with what the company is doing to help the community with their personal and professional missions. He talks about the evolution of computer technology from the nascent days of the 1890s through today, pausing to highlight Adobe as a company which was revolutionary in merging the creative with the technical and citing drafting as an industry which was transformed by the introduction of technology.

“Creative software has historically required expertise, but now new creative tools… bring their power more people , and already we can see amazing results pour out when you help people think creatively.”

Although art and analytics may seem different, Chabot identifies a common thread: both are on a mission to reveal truth and impart meaning, both challenge their viewers to look at the world through a different lens, both rely on observation and curiosity and encourage creative problem-solving.

Although the business intelligence world may not acknowledge “supporting creative thinking” as a crucial driving requirement in a RFP for purchase of a new , but better answers and ideas are certainly key to more effective analysis. As part of that goal of finding better answers, deeper and clearer thinking may help isolate trends which are invisible in a surface analysis, and the powers to experiment, to iterate quickly, and to explore are key in discovering those otherwise unnoticed elements.

Traditional BI software, though, can stifle the creative vision and make the process of quick experimentation impossible. Tableau aims to bring users the “ability to flow with data” through empowering people in four distinct channels: Experimentation, Speed, Expressiveness, and Control – and Chabot postulates that these should be crucial design requirements for 21st century business intelligence systems, on their mission to empower analysts to serve brilliant people.

Being able to prototype quickly and flexibly empowers experimentation. Not necessarily in the exact way of a Frank Gehry or Andrew Wyeth, but along the same vein allowing for ideas to be tested, refined, adapted – and resulting in comparable breakthroughs. But empowering experimentation itself isn’t enough.

“Imagine as an artist, if you made a mark on paper… and it showed up ten minutes later.”

The artist needs to be able to stay in the flow of ideas, and that continuity is important.

“How long has it been that the business intelligence software world has been dominated by wizards and flowcharts?” The mere result of such lockstep workflows is that expressiveness is stifled, often in ways which mask answers which can be otherwise brought out when limitations restricting expressiveness are removed. That removal of limitations, in turn, empowers the analyst by enhancing the control over what otherwise are lifeless crosstabs of numbers, rather than colorful and visually dynamic – and the result is visible in market sectors as diverse as air travel, game development, and sales management.

He yields center stage to co-founder and Chief Development Officer Dr. Chris Stolte, who shares the limelight with members of the corporate leadership highlighting seven areas of focus: Visual Analytics, Performance, Data Preparation, Storytelling, Enterprise, Cloud, and Mobile.

Stolte uses Tableau Desktop, as currently in development, to display some of the features which will make their debut in Tableau 9. No longer will an analyst be constrained by dragging and dropping pills on shelves, showcasing an ability to type in dimension and measure names right on the shelves themselves, with the software automatically completing the field names. Not only can that happen, but calculations can be authored in the same manner. Continuing to stay in the flow of analysis, he demonstrates blending free-form calculations by joining data on the fly by switching to display a new data source, and adding a measure from it into the existing calculation, and then saving that data model.

He goes on to show a new Analytics pane which will join the Data pane, making it easier than ever to add summary and predictive elements – incorporating them into the “Tableau Zen” of drag and drop – and then shows selection-determinant reference lines updating independently from those of a group in a view. This dynamic updating extends into a revised Table Calculation workflow – where the calculation shows exactly what a change will do in a view while the calculation itself is being authored – and without pulling the analyst out of the flow of the creative thought stream.

Extending the dynamic to another new element, Stolte shows Geographic Search – taking earthquake data with just latitude and longitude and focusing it on the fly to specific locations by simply typing in the names. It doesn’t stop there, but extends to radial selection and free-form lasso select as well, bringing new power to the user, all without forcing an exit from the investigation of the data to experiment.

Vice President of Product Development Andrew Beers takes his turn next, speaking about Performance. He shows breakthroughs in the Viz Engine, flying through a visualized painting, seamlessly panning, zooming, and selecting. Switching into data more likely to be representative, he shows a live connection to a database of 173 million taxi rides in New York City, and dives in showing results easily four times faster than the same calculations run on the same machine running Tableau 8.2, leveraging a change in the Data Engine to use multiple CPU cores in parallel. Database queries will likewise be parallelized, with cited examples reducing 15 second queries to three, and more complex ones reduced from nearly a minute to as little as seven seconds. Tooltips have been enhanced, now updating immediately as the user mouses from one area to another. Not to be outdone, mapping has received a similar boost, with new map tiles and their related data being rendered on the fly as the map itself pans. This analytic speed scales, extending onto server, allowing the speed shown elsewhere to extend even to multiple users interacting with the same workbook, in the same seamless, quick method as on desktop.

Senior Director of Product Management Marc Reuter came up next, talking about advancements in Data Preparation, and showcasing continued enhancements to the automatic data modeling introduced in Tableau 8.2, breaking three data elements combined into one field and splitting them dynamically by identifying the delimiters common to the data in the field, and then adding them individually to a view. This Split function isn't just limited to the preview, but extends to the view pane as well.

Data cleanup has been extended, and now features an automatic way to remove headers, footers, and an annoying mish-mash of merged cells and other formatting nightmares; and the data itself can now be un-pivoted and reshaped on the fly within the data preview. This should obviate the need to repeat an annoying data cleaning operation every time the data refreshes.

Data connectors have gotten some love as well, with extensions into web data connectors, empowering users with even modest Javascript experience to write simple scripts to unlock data on the web which was previously inaccessible.

Storytelling was next, with Vice-President of Visual Analysis Dr. Jock Mackinley showing off updates to the 8.2 introduction of Story Points, extending the functionality of the feature, and creating an story on-the-fly in front of the audience, pointing out the newly added thumbnail view which will debut in Tableau 9, and which should free the author from needing to remember which sheet contains which data and enabling a quick display which makes the choice of the correct view simple and quick.

Visually, changes within Storytelling will allow authors to blend colors and fonts customized to the overall look and feel and dynamic of the story - supporting it, rather than potentially distracting from it.

With the reach of Tableau into ever larger enterprises, Senior Director of Product Management Francois Ajenstat stepped onstage to talk about some of the forthcoming changes to make adoption within the Enterprise - whether internally or in external-facing customer portals - easier, more secure, and more effective than ever. Tableau 8.3 will add authentication through Kerberos and via smart cards. Server will be more extensible, as Javascript, Data Extract and REST APIs will be added in the Tableau 9.0 timeframe, empowering IT professionals to manage and deploy Tableau Server even in large and complex environments more quickly.

Searching, filtering, sorting, all will be made faster and more responsive, and added 'breadcrumbs' will help users orient themselves after searching for a particular factor of interest - potentially aiding the user in finding other workbooks which may be of interest.
Permissions will be refactored, too, making administration easier, even across complex user bases, with a new visual paradigm replacing the occasionally complicated and confusing permission model currently in place, where it can be challenging to ensure that users in groups have the proper levels of access by empowering a simple see-and-click model.

Vice-President of Product Marketing Ellie Fields took her turn next in the showcase, talking about new developments forthcoming in Cloud computing with Tableau, featuring Tableau Online as a platform for connecting Cloud data warehouses and applications in conjunction with on-premise data which can be presented in web browsers, on mobile devices, or even encapsulated in embedded applications.

Dave Story, Vice President of Mobile Products closed out the keynote demonstrations showing Tableau Mobile - faster than ever, and enhancing mobile editing by empowering calculations to mobile and web-based devices for the first time in the forthcoming Tableau 9 with a presentation model almost identical to the familiar method in Tableau Desktop - and notes that no longer will features debut in Desktop and later be added to Mobile, but rather that design and development will bring new features to both concurrently - so the changes which apply to Desktop will also be mirrored in Mobile, rather than coming in some indefinite future release. Tableau Mobile will allow users to maintain 'snapshots' of their favorite vizzes as well, even making them available with user-defined synched data offline.

He continued to debut a new product, codenamed "Project Elastic", demonstrating touch-based filtering, category selection, empowering dynamic changes of a view on the fly via a series of right- or left-swipes, automatically re-filtered in a context-appropriate manner. Elastic will be made available shortly in a pre-release state at tableausoftware.com/be-elastic.

What were the elements of the presentations which resonated most with you, or are you looking most forward to incorporating into you daily workflow? Share with us - and the rest of the Tableau Community.