Tiki RSS feed for weblogs Last posts to weblogs. http://vacommunity.org/img/tiki/Tiki_WCG.png 2012-05-19T03:33:35+00:00 Zend_Feed_Writer http://vacommunity.org/ <![CDATA[Is Visual Analtyics Software Meeting User Needs?]]>
Current visual analytic tools tend to feel complex. There are many options, choices, and avenues open for investigation. This results in a heavy learning curve with the requirement of investing significant time and effort before receiving the promised benefits. It was pointed out that this is in many respects counter to the way ‘new’ information technology is supposed to be. Web search engines and smart phone interfaces have all driven for simple, easy, and fast. We should have to do less and still get more.

This is not the case with many visual analytic tools. Where does an information analyst even start with a visual analytics tool? Which visualization is the right one? Once you get started how do you know when you are done? How do you measure progress? It is too easy to get lost in the data, to just swim around in circles. Great exercise but you might not get anywhere. That is a problem given that analysts are usually under significant time restrictions to provide meaningful and actionable assessments. These and other great examples were highlighted by Alan.

We, as an R&D community, are working hard to address these limitations. Many of which I attribute to the youth of visual analytics and the complexity of the task being addressed. It is easier for developers to leave all the options and parameter selections up to the user, giving the user complete control. As application developers, we don’t yet really understand the needs or operational and cognitive spaces users reside in. In the early days of developing IN-SPIRE(tm), many parameter selections were made available to the user. For instance, the user could select the number of clusters that documents would be binned into. This is a simple yet valuable example. The typical user did not have the knowledge or interest to play with optimization of clusters. Providing a default number of cluster and moving cluster adjustments to a secondary menu was good but not a complete answer. This resulted in menu bloat, making the tool appear overly complex and causing confusion. It then becomes too easy to start down a path of endless parameter permutations and ignore other non-software related avenues of investigation. Currently there is no option in IN-SPIRE for adjusting the number of clusters. Is that appropriate for the long term? Will users become more comfortable with adjusting options like these? Will developers become better at simplifying yet not minimizing functionality?

We have come a long way. We still have a long way to go. Perhaps I have too great of a fondness for these types of questions but addressing challenges like this is at the heart of progressing visual analytics.

]]> 2011-05-17T16:59:00+00:00 2011-05-17T16:59:00+00:00 http://vacommunity.org/blogpost22 <![CDATA[VAST Keynote Presentation]]> http://amaral-lab.org/) gave the keynote at this year’s VAST. He provided several key points. One was the importance of progressive disclosure given the mass amount of information now available in many domains. The topic of progressive disclosure is generally well understood and has been for many years but probably is not used to the degree it should be. Luis used maps as an example of this. Without progressive disclosure we would not be able to maintain a big picture when ‘pulled out’ because the image would become too cluttered. I would propose the work in level of detail for z depth rendering is another good example. In many visualizations, we try to show everything regardless of how much information is there. He points out that a lot of work is needed to better provide progressive disclosure in many domains.

From the talk and audience questions it seems like there are two aspects to the challenge, model of the data space and understanding the task at hand. We (we being those creating the visual analytic tools) need to better understand the model of the information space. Without a representational model of the information space we can’t develop a hierarchical structure to provide levels of detail. Just as importantly, we have to understand the task being accomplished. Ben Shneiderman pointed out that in the map example Luis used the task to “drive from Boston to Chicago” it wasn’t to show the pollution levels or population density. That determined the information to be displayed and what was important at each level of disclosure. This same understanding of task is required in each domain.

The chairs selected a great keynote presenter in Luis. There many other points brought out in his talk and he gave the audience a lot to think about.

]]> 2010-10-25T16:12:00+00:00 2010-10-25T16:12:00+00:00 http://vacommunity.org/blogpost21 <![CDATA[A Few of the Things at VisWeek]]>
See everyone there!


VAST Keynote Speaker on Monday

How Will Big Pictures Emerge From a Sea of ... Data?
Luis A. N. Amaral
Professor
Northwestern University

Jim Thomas Wine Sip on Tuesday

Starting at 7:15pm. Raise a glass in memory of Jim Thomas.

Birds of a Feather Session at VisWeek on Thursday

Getting to “a-HA!” - and Knowing How We Got There: A Birds-of-a-Feather Proposal for Visualization Evaluation"
THURSDAY 6-7:30PM IMPERIAL BALLROOM A


]]> 2010-10-22T21:22:00+00:00 2010-10-22T21:22:00+00:00 http://vacommunity.org/blogpost20 <![CDATA[Birds of a Feather Session at VisWeek]]> THURSDAY 6-7:30PM IMPERIAL BALLROOM A
Hope to see additional postings on this BOF!


]]> 2010-10-20T17:39:00+00:00 2010-10-20T17:39:00+00:00 http://vacommunity.org/blogpost19 <![CDATA[VACCINE Annual Meeting]]> DHS VACCINE Center of Excellence. VACCINE has been doing great work over the last year to support homeland security professionals. Purdue and many of the universities in VACCINE have been doing work in this area for many years and this has enabled the center to make dramatic progress during its first year. The VACCINE network has continued to grow and incorporate a wider group of members. Their center includes not only universities but companies and practitioner partners.

Dave Kasik from Boeing presented an interesting industry perspective looking at how his company is using and views the role of visual analytics. Boeing is highly interested in how to better understand information from a myriad of sources to improve airline safety. He gave an example where results from research they funded resulted in the way pilots are now trained to deal with bird strikes. The point I took away from Dave’s talk is that visual analytics is an important capability in addressing the problems they face. To do this effectively you have to work closely with the practitioners. They are the human-in-the-loop of visual analytics.

Rear Admiral Michael Parks introduced some of the unique challenges faced by the Coast Guard around the Great Lakes region. When I think of Coast Guard rescue, I usually think about dealing with boater safety and interdiction of boats. But given their location, they have to also deal with a large surface area of ‘hard water’ (read frozen) for a significant number of weeks. He started the talk with a lot of what they needed is “fill in the blank” for a data visualization. However, he quickly then switched to talking about the analytical needs they have and how VACCINE has helped them out. That was highly encouraging to me. The Coast Guard, at least Rear Admiral Parks, has seen the importance of moving to more powerful analytical tools not just visualization of raw data.

]]> 2010-10-06T13:37:00+00:00 2010-10-06T13:37:00+00:00 http://vacommunity.org/blogpost18 <![CDATA[VAC Consortium and VAST]]> website says) the first international conference dedicated to advances in Visual Analytics Science and Technology. The VAST program is strongly focused on presenting the latest advances in visual analytics through peer reviewed paper presentations and discussions. The VAST technical program endeavors to represent the diversity of the visual analytics technical community and each year attempts to broaden even further. As part of the larger VisWeek venue, VAST attendees have several great workshops and tutorials that are also available to them.

The VAC Consortium program is much less focused on technical advancement and more on the business of visual analytics. The program is made up of panels and speakers that provide an end user perspective, present the needs of a new community that has not been represented before, or views from industrial member representatives as examples. The consortium provides an opportunity for networking and interaction between research, business, government, and user community members. Many of the attendees do not have a technical background in visual analytics. For this reason, sessions are organized in such a way as to expose first time attendees to the field of visual analytics and to raise awareness. The hope is that many of those attending the VAC Consortium will also then attend VAST to understand the state of the technology.

]]> 2010-09-08T04:53:00+00:00 2010-09-08T04:53:00+00:00 http://vacommunity.org/blogpost17 <![CDATA[Information visualization - Beautiful vs functional]]>
I was happy to hear Irene Ros from IBM of Many Eyes fame, turn and challenge the earlier statement by saying that aesthetics are equally important. Its interesting that there are two sides to this argument. Why does it have to be one or the other?

I liken the argument to the difference between information architects and graphic designers. When did this become two separate roles? As a UX professional in the field, I do both...at the same time, I have a hard time differentiating the two. Functional and beautiful are two sides of the same coin not just in visualization but all across product, software and web design.

For the record, I whole heartedly agree with Irene here...Information visualization should be aesthetically pleasing as well as functional not to the detriment of either. There should be a balance that makes the work usable, meaningful and still appealing to use...

Thoughts?
]]> 2010-09-01T15:00:00+00:00 2010-09-01T15:00:00+00:00 http://vacommunity.org/blogpost16 <![CDATA[Day 2: New Applications for Visual Analytics]]>
Joe brings up a great point in broadening the business base for visual analytics. What impact does the VAC have in the private sector. I know from personal experience that large enterprise companies are investing, either in purchasing companies, product or developing it. CRM systems are a great example...Both Microsoft and SAP are developing aggressively. Business Intelligence is embedded in traditionally non viz suites like Sharepoint. Salesforce.com uses visual analytics in its platform. Excel is still a driving platform for data and visualization in large scale business reporting. Why is that? There is no real analysis in Excel without additional plugins, sure you can pivot etc but it doesn't typically tell the financial analyst anymore than they already know. Finance lives and dies by its use of information visualization.

How should we engage the private sector?

What is the difference between business intelligence and visual analytics?

Top investors of IT in 2010 (according to Gartner)

Pharmaceuticals
Banking and Finance
Insurance

How do we engage business in these areas in the Visual Analytic Community?

Joe recommends a comprehensive, longer terms analysis of the place for and potential value of visualization and analytics in the non-government business space needs to be performed - let's do this right. The consortium as a group should describe where and why its transformational.


]]> 2010-09-01T12:43:00+00:00 2010-09-01T12:43:00+00:00 http://vacommunity.org/blogpost15 <![CDATA[Panel Discussion: Investing in visual analytics]]> Panelists: Emily Salsgiver - moderator
Larry Rosenblum
David Kasik
John David Miller
Debbie Gracio

What can visual analytics contribute to meet the goals of government and industry?

Discussion, questions and comments

Here are some high level notes I captured during the session.

There is an opportunity in researching the marriage of predictive analytics and visual analytics. The two are not completely dissimilar but they are different processes. However, one could and should support the other.

There is an expanding volume of data...how do you synthesize and visualize? This is seen across the panel and a critical problem.

Technology transfer can be a real issue when it comes to adopting visual analytic processes, techniques and platforms. Industry is plagued with data locked in silos. The cultural bias also hinders adoption..."my domain, i must protect it".

How can we apply existing applications with real data and real problems? Research community needs to figure out how to accelerate connections with the data. Data in real life is ugly and malformed and it will not get clean, the research community needs to figure out how to deal with it.

What can we do now that will help the systems 5 to 10 years out, be better systems? What are the longer term research agendas that will impact our analysis 10 years out? The field needs to emerge as a multi-disciplinary field looking at the problems from different areas.

How can industry, government and academic disciplines work together as a cohesive community? The key will be understanding the real problems and applying new technologies and research from the community to continue to advance visual analytics. Cooperation will be hindered by cultural, IP and competition from industry. They will be reluctant to share data. However, if we can work from a shared research agenda, we can all work on the same topics.

Are we doing enough to prepare the next generation to take on these challenges? The curriculum today are not teaching analytic methods. That discipline needs to be institutionalized. Its a multi-disciplinary game, students need to be educated in a more multi-disciplinary way.

Another interesting panel...not as passionate but still a great discussion that hopefully will generate other conversations through out the day.

]]> 2010-08-31T17:52:00+00:00 2010-08-31T17:52:00+00:00 http://vacommunity.org/blogpost14 <![CDATA[VACommunity web site]]> ]]> 2010-08-31T17:22:00+00:00 2010-08-31T17:22:00+00:00 http://vacommunity.org/blogpost13