Point: Data Visualization Calls for Specialization


Point / Counterpoint, Independent Thought Leader

Park is a trusted advisor in the fields of analytics, social networking, unified communications, and enterprise mobility. His work includes insights on telepresence robotics, the true cost of Bring Your Own Device (BYOD), and the alignment between sports analytics (made popular in Moneyball) and business environments. Park is a top 10 Big Data, analytics, and mobility influencer who has been quoted in USA Today, the Los Angeles Times, and a wide variety of industry media publications. 

In addition to his analytics, telecom, and industry background, he has also been involved with online social media and social software for over 15 years, leading to a unique perspective on the social enterprise, social marketing, gamification, hyperlocalization, and the power of network effects associated with technology adoption.

Hyoun holds a Bachelor of Arts degree in Women's and Gender Studies from Amherst College and a Masters of Business Administration degree from Boston University.

Counterpoint: Train for Data Visualization Skills

Chances are you've already got good data visualization experts on staff, even if you don't know it yet.

Customer-Centric Banking Analytics Scares Me

Banks have tons of customer data at their disposal; unfortunately not all will use it scrupulously.


balance
  • 3/19/2013 8:17:02 AM
NO RATINGS

Nice post. Being able to interpret statistics and data is essential, but so is being able to communicate the results. If one person can do both, then so be it. Ideally, data scientists and professionals have both clubs in the bag.

Re: balance
  • 3/19/2013 9:58:47 AM
NO RATINGS

How challenging is it to find someone with the brain of an Analyst and the soul of an artist?

Re: balance
  • 3/19/2013 11:29:52 AM
NO RATINGS

I would venture to guess that it is extremely difficult.

Re: balance
  • 3/19/2013 6:25:47 PM
NO RATINGS

"How challenging is it to find someone with the brain of an Analyst and the soul of an artist?"

 

It's not easy.  I've been trying to balance that line for decades as a women's studies major, studio musician and a former DBA turned industry analyst, but I'd be the first to admit that my skills in data are more towards the analytical side than the visual side.  Although I can quickly identify key data points that are important, I've never been very good at creating immersive visualizations, which is part of why I'm convinced that data analysis and effective visualizations are two separate skills.  The few who can do both are worth their weight in gold. 

Re: balance
  • 3/19/2013 6:57:20 PM
NO RATINGS

What an interesting background you have Hyoun

Re: balance
  • 3/19/2013 9:18:07 PM
NO RATINGS

@Hyoun, do you agree, though, that every data scientist should have solid instructions in visualization and communication skills, and given their fair shot at becoming visualization masters? In other words, though analysts who are also master visualists are rare, it doesn't hurt to turn over every stone.

Re: balance
  • 3/19/2013 10:26:23 PM
NO RATINGS

To be honest I am really struggling with this concept of data visualization, I can understand how it is beneficial to those who don't want to be bothered with the details of data but how can you trust the visualization is valid ? I guess we do this all the time in one way or another, but from analyst perspective I dobn't understand the argument.

But more to Hyoun's point, I don't think we need to make a special title for this function because it is something someone in a team will have some ability to carry out.

Re: balance
  • 3/20/2013 8:15:14 AM
NO RATINGS

Hi Louis. I think you have a valid question, but one that applies not just to the visualization of analytical results but analytical results overall. That points to the importance of making sure data goes through all the front-end work to make sure it's clean and the results are truthful. Can a poorly designed visualization lead to misinterpretation of the data, absolutely. That's what's nice about analytics tools that render the visualizations automatically -- you still need all the expertise to know what data to input and how to spot patterns, but at least the design bias disappears.

Re: balance
  • 3/20/2013 11:37:02 PM
NO RATINGS

Hi Beth,  Thanks for your excellent explanation, it really helps me digest what seems so difficult for me to wrap my head around. I am going to really think about your great points on this topic.

Re: balance
  • 3/21/2013 8:48:43 AM
NO RATINGS

Glad I could help.

Re: balance
  • 3/20/2013 10:47:22 AM
NO RATINGS

"I can understand how it is beneficial to those who don't want to be bothered with the details of data but how can you trust the visualization is valid ?"

@Louis, we all put bias into our analysis.  Sometimes the bias comes from our data sources.  Sometimes it comes from our data inputs or data queries.  And sometimes it comes from additional drill-down analysis or correlation analysis.  We make assumptions about the key metrics and data that have downstream consequences. 

Visualization is another area where bias, judgement, and skill come into play. Yes, someone in the team will ultimately have some ability to do this or be assigned to do this.  But will that visualization be good enough to engage line of business users?  Or will that person leave out a visual detail that makes a key difference in understanding the results? Is that person thinking of time-series animation of data? Or creating analytics-based customer categorizations? Or simplifying social network analysis diagrams to identify core linkages and relationships?


The status quo is non-specialization and, honestly, that's where the vast majority of companies are today.  But the value-add opportunity comes from improved visualization and discovery.

Re: balance
  • 3/20/2013 11:43:12 PM
NO RATINGS

@Hyoun   Well said and you poise questions that I had not considered for certain. This bias does happen with all analysis so if data visualization can migitigate this to a certain degree, then I can see the value in it's use.  And I like your final point best, most companies do not consider this, so if you can tailor data effectively in this case in terms of visualization, I agree it does seperate you from the pack.

Re: balance
  • 3/25/2013 2:30:02 PM
NO RATINGS

"Visualization is another area where bias, judgement, and skill come into play."

Visualization does deliver some value to the data and the business, even though it might be deceptive sometimes. But as goes the saying, "a picture is worth a thousand words".

Re: balance
  • 3/20/2013 8:10:56 AM
NO RATINGS

Good question Broadway. One I'd like to know the answer on as well!

Re: balance
  • 3/20/2013 10:39:41 AM
NO RATINGS

"do you agree, though, that every data scientist should have solid instructions in visualization and communication skills, and given their fair shot at becoming visualization masters?"

@Broadway, this makes sense.  It's easier to train data scientists and data analysts best practices in visualization than to start from scratch.  But I think it's also realistic to assume that many data analysts are going to be challenged to internalize these lessons.  It doesn't mean that they're not useful or valuable; data crunching is still a rare and valuable skill.  So, train your existing staff, but don't crucify the ones who don't get it after putting in a reasonable effort.

Re: balance
  • 3/19/2013 10:27:56 PM
NO RATINGS

@Noreen   I want to say it is not that difficult, but maybe it is.

Re: balance
  • 3/19/2013 4:12:58 PM
NO RATINGS

Seems to me that the good analytical professionals have always been able to tell a story with the data or about the data. But the visual element adds a whole other dimension to that!

Re: balance
  • 4/16/2013 3:34:41 PM
NO RATINGS

I appreciate the passionate argument, but it only holds in a small corner of the problem space. Data visualization is a surprisingly old subject, and it's been given a thorough treatment by a long list of distinguished researchers. The good news is that, although advanced research continues, the day to day needs of the average jane/joe are covered. When it comes time to present data, we have ridiculously powerful tools now that are literally under our noses, and they will handle the majority of our needs right out of the box. For the record, I count 74 different chart types in Excel. Seriously, Excel!

Occasionally -- I would even say rarely -- we may find it necessary to create more complex or specialized representations; the good news here is that self-education in this Internet age is all too simple. Go read a book by an expert like Tufte. If that's not enough, you can take courses from some of the masters without even leaving your office, for pete's sake. 

Let's skip over the bit about artists and left brain vs right brain (complete myth, by the way: http://www.psychologytoday.com/blog/brain-myths/201206/why-the-left-brain-right-brain-myth-will-probably-never-die). At bottom, it is obviously true that people have different strengths and interests. However, most of our daily work presenting facts and figures will not require exceptional ability. Sometimes, again rarely, a breakthrough visualization method means the difference between communication and confusion, c.f. Feynman Diagrams. Most of the time, almost all of the time, that's just not the case. 

Here's the secret to effective data presentation, and it's been rediscovered so often it's not even funny. Keep it as simple as possible. And to be honest, it doesn't take an expert to do that.

Re: balance
  • 4/17/2013 10:01:06 AM
NO RATINGS

@jmarkdavis, thanks for jumping into this conversation. I don't disagree with your points but I do wonder if they're oversimplified when put into a big-data context -- especially if you're talking about using visual analytics as a way to better explore that data (so less of a means of presenting results). 

Re: balance
  • 4/17/2013 11:09:33 AM
NO RATINGS

I understand what you're suggesting, but think about the practicality of it. Simplification is precisely the point. If the analytics and big data tools available in a given enterprise aren't getting the job done, what is the most efficient and effective remedy? I can tell you from years of first hand experience that seeking or creating an expert capable of inventing and then building something entirely new is a losing proposition, unless of course that enterprise happens to be in the business of building and selling such tools. The smart money will conduct a product search, find a product with suitable exploration features, acquire that product, train appropriate roles, and quickly reap the benefits.

Re: balance
  • 4/17/2013 11:12:42 AM
NO RATINGS

jmarkdavis, Can you share an example of what you consider a good visualization?

Re: balance
  • 4/17/2013 12:29:46 PM
NO RATINGS

I like the idea of an example, Noreen, but "good" depends on the need. Some that I find visually appealing, just because they look cool and are fun to click through, are treemap, tag cloud, and fish-eye. My question -- consider it rhetorical -- is just what sort of data can't be explored using the wealth of visualization methods we have on hand today?

Re: balance
  • 4/17/2013 12:17:22 PM
NO RATINGS

@jmarkdavis, so, in a nutshell, it's all about finding the right tool?

Re: balance
  • 4/17/2013 12:38:31 PM
NO RATINGS

Yes, Beth, exactly. Thank you. In the vast majority of cases, there's a tool somewhere that will satisfy. Ten to twenty years ago, it might have made sense to assign a company resource to data visualization. Today, enough very smart people have made data visualization their full time job, and produced commercial quality products for visualization, that experts and specialists are no longer necessary. As for artful presentation, big companies have communications or art departments when that sort of thing is truly necessary. 

Visual art
  • 3/20/2013 3:35:02 AM
NO RATINGS

I think a data scientist should be good in: Computer Science (coding)/Big data analysis/Graphic design/and Human-Computer interraction (HCI). This should be the data superman the world can ever have. I find it strange if you say "a traditional data scientist" yet even this name has not picked up yet, meaning that the field is still new and exciting! Email: aggreymutimba@yahoo.co.uk

Re: Visual art
  • 3/20/2013 6:59:11 AM
NO RATINGS

Geez Makokha!!! Not asking for too much, are you?

Re: Visual art
  • 3/20/2013 6:59:36 AM
NO RATINGS

Btw--welcome to our community

Re: Visual art
  • 3/20/2013 7:30:26 AM
NO RATINGS

:-)

INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +