Held under the “Making Data Work” motto, this year’s edition of Strata Conf. in Santa Clara, California, brought together the leading minds in Big Data who are using it to drive business strategies, as well as the practitioners who collect, analyze and manipulate it.

Joe Hellerstein, Founder & CEO with Trifacta and a Professor at UC Berkeley, took the stage to elaborate on the challenges posed by Big Data and the analogy between most analytic effort and ground control.

A Chancellor’s Professor of Computer Science, Hellerstein’s work focuses on data-centric systems and the way they drive computing.

“I’ve been working with data for 25 years and I’ve never experienced a time like this,” confessed Hellerstein. “The data people are transforming computation, technology and society.”

Included in 2010 in Fortune Magazine’s list of “50 smartest people in technology”, Hellerstein seems convinced “this is the time for Big Data moonshots.” He went on to speak in depth about the persistent mythology in the community and the press that data scientists are heroes carrying three superpowers: moduling, coding and visualizing. He wanted to debunk the myth, specifying that “eighty percent of the Big Data work represents cleaning the data” – an actual quote of DJ Patil.

Despite being an old and persistent problem and despite many attempts, “automation is not the solution to data transformation and data cleaning,” stated Hellerstein. “Achieving liftoff takes people.”

The problem fundamentally lies with human judgement on taking particular data sets and transforming them for specific use-cases. In the end, any Big Data project is going to succeed or fail based on the empowering a wide variety skilled people to get the job and the task off the ground.

The “Big Data Moonshots and Ground Control” presentation included two perspectives – in tech and design. To talk more about the design part, Joe Hellerstein invited on stage Tutti Taygerly, VP of User Experience with Trifacta.

“Enterprise software is not traditionally known for being well-designed. We believe it needs to change its focus on the user,” said Taygerly.

Continuing the theme of ground control and space conquering, Taygerly showed a slide presenting the audience many of the information factors the astronauts have to take into account, almost simultaneously, accurately pointing out that “the burden to not mess things up in this environment is entirely on the user.”

Prompting a trip down memory lane, Taygerly highlighted that “looking in the data transformation space, historically it hasn’t been a lot better.”

The command line was super powerful as long as the user knew what to do with it – this put pressure on the user. In time things got prettier, but not necessarily better. Even something that blurs the intersection between consumer and business app, like Excel, still puts the bulk of the burden on the user. You have to select a particular task, you have to know what macros to write.

“We believe, in designing software, that it’s time to remove the burden on the user,” announced Taygerly.

The futuristic world depicted in the movie “Her” shows the intersection between the machine and human is so seamless that the human can fall in love with his operating system. The continuous learning process in the movie puts the burden on the machine, not the human. In comparison, Tom Cruise’s world in Mission Impossible highlights the same burden on the user.

Taygerly continued: “At Trifacta we rely on the course of three design principles:

 .

1. Visually simple, Functionally strong – when you look at a set of data, we believe there a two things that you want to see. a) the actual data value; b) the simple visual summary of what’s happening. Rather than overwhelming the user with lots of pretty visualizations, we want to do one universally-understood visual. Letting you move from raw data to simple histograms that bind the data based on context, type and values.
2. Elegance at Scale – we let you focus on one single data value and let the simple amplify, magnify and work at scale, allowing seamless movement between micro and macro.
   
3. User Empowerment – humans are naturally emotional, fallable creatures. We allow for exploration and discovery in our data sets. We allow for mistakes and failures with no penalty. We allow iteration and reiteration until we can celebrate success with you.

 .

In addition to these three design principles we rely on a set of user persona against which we validate our designs.”

 .

DATA SCIENTIST:
“I want to work on predictive analytics to solve fraud and customer retention problems for the bank and not waste my time with data prep.”

BUSINESS ANALYST:
“I need direct access to all the data so I can rapidly turn around data-driven insights to inform business decisions.”

IT PROGRAMMER:
“I am here to help the organization. I can help pull data and setup data marts. Overall I’d love to see processes become more efficient and self-service.”

 .

Hellerstein returned on stage to exemplify the beauty of Trifacta with a demo.

“This company is rooted in the research we’re doing at Berkeley and Standford that led to this interaction model which enables this technology; we call it Predictive Interaction. It begins by user interacting with data visualizations on the raw data. As they do that, algorithms built by Trifacta are predicting what actions the user might want to take, based on the data that’s in the system and based on the context of what the user has been previously doing.”

The visualizations are not static. You can interact with them and, as you do, the system predicts transformations that you might want to do to the data. You can overview what the transformations will do to the data with previews. In addition to interacting with visualizations, you can interact with anomaly detection, see errors in the data and receive recommendation on how to remediate those issues.

“The system learns from your behavior and you learn from system preview and suggestions,” clarified Hellerstein.

Living up to Trifacta’s slogan  – People transforming data – Joe Hellerstein concluded his presentation and demo with the following thoughts: “People who are working with data are the focus of what the community is working on. Data transformation is one of the key challenges that we can make both agile and fun.”