Last week, we hosted a Tweet Chat on big data with several analysts. And we thought the chat was cool enough to highlight on this beautiful Friday morning. So grab a cup of Joe and see what you might have missed in the #SFBigData Tweet Chat.
Before we get to the meaty Q & A, let me point out a few highlights:
- This Tweet Chat had 492,596 impressions on Twitter and reached 43,960 Twitter accounts.
- Shawn Rogers (@shawnrog) from EMA Research received the most retweets and he may have coined a term – Hadump. But you’ll have to read on to find out what it means.
- Our list of contributors includes 26 individuals, with Shawn Rogers and Ted Cuzillo (@datadoodle) leading the pack in number of tweets. Other players include Cindi Howson (@biscorecard) from BISorecard, Neil Raden (@neilraden) of Hired Brains and Jorge Garcia (@jpgtec) of TEC. You can find the complete list of contributors in our full tweet chat report.
The Q & A
This is not a complete transcript, but what we hope to give you is a good overview of the conversations and insights.
Question 1: How would you define #BigData in 140 characters or less?
Rogers: Big data = data sets that can no longer be easily managed/analyzed w/traditional/common DM tools, methods & infras.
Rogers also suggests that the term big data “evolved to be more about the opportunity and analytics than the description of data.”
Cuzillo suggests that big data has a variety of definitions. He tweeted: “As that Sup Ct justice said, you know it when you see it.”
Question 2: Is big data the future or is it all hype?
Rogers: “The hype is still present but real world, valuable use cases are becoming more commonplace.”
Cuzillo: “The part that’s hype is so because in so many articles, etc. BD has become synonymous with ‘data.’ But BD’s much, uh, bigger.”
Rogers says that analysts have seen much higher adoption than they expected at this point. And Cuzillo compared big data to television, saying that at first it wasn’t much better than the radio. However, it expanded with “resolution, dimensions and applications.”
Question 3: In your conversations with enterprises, what is the #1 big data question or concern you’re hearing?
Rogers receives the most retweeted tweet of the chat with this answer – “I’m tired of the 3V’s. It’s about complexity, workload scale and speed.”
Cuzillo says that he’s hearing questions about who does the “making sense” work of big data and “how can insight from down the chain be sent back upward?”
A Side Discussion on Complexity
From Brett Stupakevich (@Brett2point0) of Spotfire, Rogers and Howson:
Howson agrees with Rogers in Question 3 that big data’s biggest concern is complexity rather than volume. She continues, “Bigger than big data is the need to mash data, makes it more complicated and then lack of common definitions.”
Stupakevich asks Howson and Rogers, “In a way doesn’t ‘complexity’ relate to the Big Data V for Variety (unstructured)?”
Howson’s reply: “Partly – variety in unstructured but u you have so much complexity w structured.”
Question 4: What key big data issue do you think companies need to pay more attention to?
Rogers: “Making sure they have a compelling business issue to solve before jumping into the Big Data world. Too many companies are chasing the “sparkly toy” without a plan. Causing #BigData investment to turn into a Hadump.”
Cuzillo: “Similar to small data: recruitment, training and management of data analysts. The whole process is just more important now.”
Howson: “Same rule for BI as big data focus on business value – what goals to align to. Environment.”
Raden: “Data doesn’t speak 4 itself. Looking only at data >false negs/pos. OK 4 hypothesis formation. Need modeling formality. Q4 #bigdata no different than #bi >informing people or process is only half the solution. Decision management needed.”
We asked Raden if a data scientist is the right conductor for the job? He says, “It depends on how you define data scientist, but in general, I’d say no. Marketing still has to do marketing.”
Question 5: What industry do you think is doing the best job of tackling the big data challenge? Why?
Rogers says financial services is by far the biggest adopter of big data, and they were willing to pay their way into early adoption.
Cuzillo says, “It’s the ones already attuned to metrics, e.g. direct mail marketing, at it for half a century already.”
Question 6: What are some of the big data analytics best practices you’ve seen?
Rogers says that mapping big data across the organization and including the business and IT stakeholders as well as integrating new data such as social data are extremely valuable practices.
Cuzillo points out that you need to know when small data does just as well and that companies need to use the appropriate tools and people for success.
Question 7: Where do you think data/discovery analytics will fit in the big data solution set?
Rogers says that data/discovery analytics is a leading use case for big data and Cuzillo asks, “Is BD worth bothering with if you don’t have data/discover analytics?”
We thought that was a cool question and take that to mean Cuzillo considers data analytics an “integral component.”
Question 8: How does predictive analytics fit into the big data picture?
Rogers: “The demand for self-serve is rising. Big Data powers deeper more accurate predictive models at speed that greatly eclipse performance just 2 years ago.”
Raden: “The real killer app for predictive modeling will be optimization. The problem with predictive models is that they can tell you what is likely to happen, not when it will.”
Cuzillo: “It’s a huge part and will be even more so as PA becomes more commonplace for non-analysts.”
Garcia: “It is more productive and efficient to turn big data a part of pred solution instead of a reactive one.”
Stupakevich: “Forrester analyst Mike Gualtieri (@mgualtieri) recently offered a “refreshing” answer to Q8 . . . “#BigData predictive #analytics is like making wine: 1) Put in big vat 2) mash it up 3) siphon the good 4) savor 5) bask 6) repeat.”
Question 9: What value does tackling the big data challenge bring to an enterprise?
Rogers: “Big Data analytics value is rooted in speed, scale & complexity. Allows the ent to do what we couldn’t 10 years ago. Early #BigData analytics value stories demonstrate great value in $$$ but also in new found flexibility.”
Cuzillo: “Resolution in the big picture, new dimensions of observation, + still-to-discovered applications. Surprises ahead.”
Question 10: What’s the biggest roadblock to big data analytics?
Rogers: “EMA Research shows #1 issue is lack of application mgt. features in big data solutions. This hurdle slowing adoption. Lack of skills to manage NoSQL environments is a close second on our list.”
Cuzillo: “A10 Scarcity of know-how, mechanisms to propagate insight too slow or awkward, as @shawnrog said, lack of app mgt.”
Garcia: “Adoption, training, maturity.”
Key takeaways:
- Big data offers opportunities for companies that invest in the talent and tools that help it contribute insights to the business.
- Big data analytics requires a road map for success and input from both business and IT stakeholders.
- Big data enhances predictive analytics, especially with social data, but it can’t predict when events will happen.
- The newness is a major factor affecting training, talent and adoption.
- Self-service big data analytics tools are the keys to success.
Next Steps:
- Subscribe to our blog to stay up to date on the latest insights and trends in big data.
- Register for the Spotfire 5 Webcast with Lou Jordano (@loujordano) on Wednesday, Oct. 17 at 1 p.m. Eastern.
Amanda Brandon
Spotfire Blogging Team
