from Rensselaer Polytechnic Institute
Web experts ask scientists to use the Web to improve understanding, sharing of their data in science
Troy, N.Y. – Peter Fox and James Hendler of Rensselaer Polytechnic Institute are calling for scientists to take a few tips from the users of the World Wide Web when presenting their data to the public and other scientists in the Feb. 11 issue of Science magazine. Fox and Hendler, both professors within the Tetherless World Research Constellation at Rensselaer, outline a new vision for the visualization of scientific data in a perspective piece titled “Changing the Equation on Scientific Data Visualization.”
As the researchers explain, visualizations provide a means to enable the understanding of complex data. The problem with the current use of visualization in the scientific community, according to Fox and Hendler, is that when visualizations are actually included by scientists, they are often an end product of research used to simply illustrate the results and are inconsistently incorporated into the entire scientific process. Their visualizations are also static and cannot be easily updated or modified when new information arises.
And as scientists create more and more data with more powerful computing systems, their ability to develop useful visualizations of that data will become more time consuming and expensive with the traditional approaches.
Fox and Hendler ask the scientific community to take some important lessons from the Web.
“…visualizations on the Web are becoming increasingly more sophisticated and interactive,” they write. At the same time, those Web-based visualization are also inexpensive and easy to use, according to Hendler and Fox.
Simple Web-based visualization tool kits allow users to easily create maps, charts, graphs, word clouds, and other custom visualizations at little to no cost and with a few clicks of a mouse. In addition, Web links and RSS feeds allow visualizations on the Web to be updated with little to no involvement from the original developer of the visualization, greatly reducing the time and cost of the effort, but also keeping it dynamic.
“Visualizations are absolutely critical to our ability to process complex data and to build better intuitions as to what is happening around us,” the researchers write. They use the example of an online weather report. With such visualizations, Web users can click on their area for a forecast or watch videos specific to their region. Without these visualizations, no one but a trained meteorologist would be able to make sense of the mess of raw data behind those pretty maps and graphical snow clouds.
In addition to the ease of using and developing visualization on the Web, visualizations on the Web can also be easily modified, updated, customized, and recreated by other users thanks to the use of Uniform Resource Identifiers. This “linking” of data is a key feature of the new vision that Fox and Hendler outline. It is of particular importance when dealing with what they refer to as “big science” on topics such as climate change that involves data that ranges from distinct fields like biology to geology.
“The challenge is that many of the major scientific problems facing our world are becoming critically linked to the interdependence and interrelatedness of data from multiple instruments, fields, and sources,” they write.
Fox and Hendler urge scientists involved in such vital scientific projects to take some tips from large Web companies like Google and Facebook, and even massive online communities such as World of Warcraft. These large companies use new data integration approach such as NoSQL, “big data,” and scalable linked data to rapidly expand and maintain their capabilities. These new capabilities provide easy-to-use, low-end tools to generate visualizations and scalable tools for curating very large visualization projects that scientists can model their own visualization after, according to Fox and Hendler.
For more information on the research of Fox and Hendler as well as the Tetherless World Research Constellation go to http://tw.rpi.edu/.
They also have a blog at: http://tw.rpi.edu/weblog/
I don’t there is anything that can help real climare, I mean climate.
While I see the obvious advantages, I also see (yet) another avenue of manipulation of the underlying data, if that data is not made easily available to those who want to see PAST the pretty visuals.
“And as scientists create more and more data with more powerful computing systems, their ability to develop useful visualizations of that data will become more time consuming and expensive with the traditional approaches.”
Maybe there’s too much ‘data’ creating’ going on and not enough ‘data recording’.
Following the RPI link reveals lofty goals and the usual academic jargon, grandiosity, buzzwords, and vagaries. Noble intentions, but no substance at this point.
“…RPI’s Tetherless World Constellation will address this emerging area of “Web Science,” focusing on the World Wide Web and it’s future use….” Oy.
As of right now, I see no applicability to the primary problem of AGW, which is, of course, the fact that it’s in the realm of fairies. Bad data will always be bad data; bad methods (no matter how many petaflops are employed) will always be bad methods.
Still, I think they should be funded with some of the money now being squandered by Michael Mann and the like. RPI TWC is far less likely to waste funds and create a dysfunctional politico-academic circularity function, d³s/dt³. The name Peter Fox rings a faint bell…
Fox is associated with NCAR and is one of the organizers of C2DC 2011 : International Workshop on Climate Change Data Challenges.
Provided that access to the RAW data is also provided, along with the capability to actually plot that same RAW data….
fancy graphics dont really impress me. long columns of numbers and detailed records and calculations for adjustments do.
If I wanted fancy graphics, I would use a gaming console.
Visualizations. That’s a double edged sword if ever there was one.
I have a side hobby, I track earthquakes and look for odd stuff just to stay on top of some of the rather STUPID things that come flying about on the web. As part of my munging around in the data, I run across charts and maps with the different earthquake plots, and usually am a bit dissatisfied with what I find. So, I dig deeper in order to find the actual quake lists so that I can use a program such as Dplot in order to wrestle the spatial info into something I can look at or skew for a better view. (Note, many thanks to A. Watts and WUWT where I discovered the program).
( A non weather/climate related quake plot of a swarm in Iceland)
http://i55.tinypic.com/16k25bn.png
The other edge of the sword, is that some people will stop providing or allowing access to the raw data for the general public. They will throw a graphic up and point to it as proof of what ever it is they are trying to say. This lack of access will hurt the amateur or the hobbyist since it will add one more brick in the wall of exclusive access.
Many of you here have seen what sort of frustration a lack of access to data can bring.
Anyway… that’s my opinion. More visualization is good.. as long as they explain what the #### it is in the plot. We aren’t morons if they can’t explain their work.
This might represent a problem. ;O)
Here’s my visualization:
Post Hypothesis on WUWT, get you academic ass kicked into next winter,
‘Nuff said,
[d]
They come across as politicians and civil servants rather than scientists.
Interesting piece… and I see they mentioned online gaming which always piques my interest. A more interesting example of a complex virtual reality would be the virtual universe that was created for EVE online:
http://www.eveonline.com/
While it only has about 300k subscriptions (players) compared to World of Warcraft (aka WoW) at 10-11 million, the difference with EVE is that they all play in one instance of the game. In WoW the player base is spread over hundreds of instanced replications of the same world (called servers or shards in other games) with only 20-25k accounts per server. In that respecct what EVE programmers have achieved is amazing. Having pitched battles with hundreds of spacecraft simultaneously must be something to behold.
EVE is a behemoth of a game in that there is a fully functioning player run marketplace, complex enough to attract research economists to study it from without. I digress somewhat…
Ironically it is through playing games that we often develop sophisticated visualisation systems, so I can see where these guys would be coming from.
For more EVE info see:
http://en.wikipedia.org/wiki/Eve_Online
If you want to see what good presentation can do, check out http://www.ptable.com/
Nothing too fancy, but all clearly laid out and easily accessible.
Of course, no presentation means anything unless the data behind it is solid and the processing of it is done the right way …
‘Maybe they could help Real Climate’
I don’t know about that, in one respect they have been spectacularly successful.
There’s one simple graphic which lingers and seems impossible to erase, as is apparent in some of the contributions here, and there’s nothing sophisticated about it:
http://joannenova.com.au/globalwarming/graphs/mwp/synthesis-report-summary-tar-hockey-stick.jpg
Don’t know if the tetherless Web Expert constellators link to this, but these are nice visualitions, eh, vilusianations:
http://www.processing.org
“DirkH says:
February 13, 2011 at 3:34 am ”
Oh, i should give this link as well, it isn’t intuitive to find from the main page.
Very fascinating artwork done with processing:
http://www.openprocessing.org/
This “linking” of data is a key feature of the new vision that Fox and Hendler outline. It is of particular importance when dealing with what they refer to as “big science” on topics such as climate change that involves data that ranges from distinct fields like biology to geology.
————-
Re: key feature of the new vision that Fox and Hendler outline.
This implies that Fox and Hendler are promoting their own new idea.
Either the press release person is incompetent or these guys are shameless self promoters.
Someone has forgotten the www was devised at CERN to support scientific research. And since then we have had the semantic web as well. And on top of that there is a lot of work in the area of data mashups via web services and metadata harvesting and publishing and data curation, etc. etc. etc……
And what has this all this to do with Real Climate?
“And as scientists create more and more data with more powerful computing systems……” What????
I always input data but never created any or got it as an output. On the other hand, I am not a “scientist”. Maybe I need a clearer understanding of “data”.
“The problem with the current use of visualization in the scientific community, according to Fox and Hendler, is that when visualizations are actually included by scientists, they are often an end product of research used to simply illustrate the results…”
Do you mean visualizations like this?
The “scientists” here were using climate visualization in a unique way – specifically…
“We wanted to find a way to bring this critical issue back into the headlines while making people laugh…”
(PS – What was Realclimate-FentonComunnication-GeorgeSoros’s position on the 10:10 “exploding children” video? Does anyone recall? Perhaps it was too “robust”…)
If the data does not speak for itself, may be there is little of significance in the data!
Real Climate and the people behind it are so far away from any reality that no help in the world could bring them back on the track.
I suggest we continue treat them as the basket case for climate alarmism.
Basically what they propose has already happened.
Scientists who want to do honest work and communicate clearly (who are fairly common outside of crimatology!) are already using the Web instead of paper. Consumers of scientific information have already learned how to sort and use Web info, and have already stopped reading paper journals, because we understand that paper journals are peer-reviewed, thus dishonest.
When one decides to read an article in some sort of a source of news, one has an expectation of gaining some information. When the writing in the particular article is so obtuse and sense-free that after several minutes, one has garnered absolutely nothing useful, one reaches the obvious conclusion that the writing is deliberately obtuse. My goodness, what a clunker. “Word cloud”? “Data visualization”? Someone should try speaking English…
When I read “And as scientists create more and more data with more powerful computing systems…” I almost decorated by screen with my morning coffee. B Shellenbaum said it well at 10:48. There’s a fundamental problem when a data analysis tool is used instead to create data. Do they realize that they’ve left the realm of science when they do that?
There’s a mixup in their paper which demonstrates the disconnection between real science and modeling.
At both Los Alamos and Lawrence Livermore Labs, there are scientific codes which can run physics experiments from “first principles.” They start with atomic components and wind up modeling things like the impact of a laser beam on a microscopic aluminum ball containing Deuterium. The model shows a shockwave propagating in the aluminum around the ball. The question was, does the shockwave exist? The answer was found in looking a high speed images captured by an Italian laboratory, and it was true, there was a shockwave. This is physics simulation from first principles, this was a breakthrough in computational physics. This simulation ran for many weeks (probably months).
The difference between that and what the climate guys are up to is that the physicists look to verify every prediction of their models, in detail. Climate science, not so much. CS tries to change the data that doesn’t fit the predictions. As was said here before – EPIC FAIL!
The visualization problem is constant, every time someone thinks they’ve solved the problem, someone else cranks the knob back up to 11, and we’re back to trying to see what is there. If you generated modeling data, then you need to see if the model reflects reality. If you’re trying to extract information from measured data, then you’re trying to figure out the best algorithmic description of the data to model the process. Either way, you need to see what you’ve got.
Personally, I don’t believe allowing some unsupervised algorithm to present the data is necessarily a great idea. I think a major part of being a scientist is knowing about the area of research, but keeping a watchful eye out for “Nature’s Little Surprises.” That doesn’t usually happen without deep involvement in the graphical representation of the data. Bob Tisdale, Willis, Anthony, and others here on this site offer real life examples in how to communicate with clear graphical imagery.
LazyTeenager says:
February 13, 2011 at 3:56 am
“Someone has forgotten the www was devised at CERN to support scientific research. And since then we have had the semantic web as well. ”
Fantasy. Tim Berners-Lee has spoken about the semantic web for 20 years now, so much is true. While XML is in widespread use now as a data exchange protocol, we are nowhere near having an ontology that would allow software to figure out the meaning of arbitrary XML data. No ontology – no semantic web. The only progress we had is that our machines have become better at guessing the meaning of stuff; which is definitely helpful, but that’s not how it was planned by the Great Visionaries.
Mike D in AB says:
February 13, 2011 at 7:48 am
“There’s a fundamental problem when a data analysis tool is used instead to create data. Do they realize that they’ve left the realm of science when they do that?”
A computer can of course be used to create data in its original form – for instance, measuring the travel times of network packets; in this case, the computer becomes the instrument. Don’t know whether the Tetherless Wizards Of The Web mean it that way, though.
“Watson” to the rescue. If it can learn to recognize voice patterns, associate them with known patterns, remember relationships, make unbiased decisions, and provide good answers to questions; it should be able to produce good visuals from historic and realtime atmospheric data that is better than our present weather models. (Much better than existing climate models.)
I’m thinking that it should be mentioned that in the context above what ‘processing’ means:
Per: http://en.wikipedia.org/wiki/Processing_(programming_language)
.
Dirk
processing is very cool. A few years back we looked at putting it on an open phone
http://www.tom-carden.co.uk/2008/11/24/processing-10/