I believe there’s an emerging space in the news industry for data visualization. Not as a supplement to the written word but as the primary source of information. Data visualization techniques, if done right, can go a long way toward addressing the bias and oversimplification that is or appears to be in many written-word news articles. Data graphics that show all of the data practically by definition have no bias. And subtle shades of gray, like how the unemployment rate varies across the nation and across employment sectors, stand out more clearly visually than they do through the written word. Yet no significant news story has been broken or even been retroactively explained via data visualization techniques. If I had only one wish for 2010 it would be:
I wish for a significant news story to break first in the data visualization space (ideally on this website).
Perhaps the biggest impediment to this wish coming true is that it suffers from a “chicken and egg” problem. Specifically the current news media is primarily verbally aligned. Some media outlets have excellent in-house graphics departments, but it appears they primarily serve in a support role to the verbally researched news stories. For a story to break it first has to be reported somewhere and then picked up by other media outlets and repeated. There are plenty of data visualization websites on the internet today so the first part of this two part formula may well happen in 2010. Where the process breaks down is in the dissemination of the story. Will verbally aligned reporters and editors understand the story if most of it is told via data graphics and not the written word? If they don’t understand it, they probably won’t repeat it in their own media. So the story languishes entirely within the data visualization space and not otherwise breaking out to the broader media. Imagine if Watergate, Iran Contra, or the Valerie Plame stories were first reported in Chinese. How long would these stories have remained in the Chinese American news media before they would have finally broken out into the broader news media?
But even if verbal reporters and editors saw a story breaking in the visualization space and understood it there is still the possibility that they wouldn’t believe it. Data visualizers examine vast amounts of information and distill it into easy to understand graphics. Historically data alone has taken a back seat in breaking news to a first person perspective. Watergate was broken by “Deep Throat” and verified by the data (follow the money), not the other way around. More recently Harry Markopolos, a financial expert, tried for years to get the SEC to shutdown Bernie Madoff, but no one took him seriously till Bernie’s own sons turned him in.
Markopolos’s inability to convince the SEC may have been due to a lack of comparables. What do the finances of a legitimate investment firm look like? What is the range of normal behavior for them? How far outside of that normal behavior were Bernie Madoff’s finances? I suspect (but do not know) that Markopolous reported only how unusual Madoff’s finances were and the people inside the SEC couldn’t easily grasp how far outside of the normal range those finances were. So for data visualization alone to be able to break a news story readers and writers may first need to get used to data visualization. This brings me to my 2nd wish for data visualization in the new year:
I want to see data visualization articles explaining the key stories of 2009 and have those picked up by the main stream media.
- Bernie Madoff. What did Harry Markopolos see and does it lend itself to visualization?
- The October 2008 TARP bailout. Everyone says the economy was about to over a cliff. Very graphic language, yet so far no compelling graphic imagery to go along with that language. I imagine lending by banks to companies could have suddenly stopped. A graphic showing that all by itself might make casual observers think bankers capriciously stopped lending, but such a system-wide decision could have occurred by a sudden collapse of those mortgage backed securities. If so, that graphic should be shown too.
- The November 2009 Breast Cancer Screening Recommendations. The panel said the risk of screenings outweighs the benefits for women under 50. This sounds like a match made in heaven for a data graphics article. Intersecting risk/reward lines. Perhaps detail based on family history which could also show under what circumstances the recommendation doesn’t apply.
- ClimateGate. There’s probably several stories here some will lend themselves to data graphics better than others. My suspicion is that those stories that can be told graphically will probably disprove the scandal and (if you believe the media only covers contentious issues) thus provide even less of a reason to cover it. One such possible graphic shows temperature data from a variety of sources all plotted on the same chart. You can see both the discrepancies between sources as well as the … crepancies?
- Health Care. Why is the US so much more expensive than the rest of the world? I would love to see a simple chart that showed where the US health care dollar goes in comparison to other industrialized countries. Is it our insurance? Are we overly litigious of our medical providers and is this a factor? Is the US population sicker than other industrialized countries? Is it because the rich are spending vastly more on health care than the middle and poor are and hence it’s the rich who are driving the “average” health care costs up? It’s sad to think that the health care bill is almost through the Senate and these answers are still not widely known.
The best experts are able to explain their reasoning to the lay public, hacks resort to “trust me, I’m an expert”. This site has many examples where data graphics show the truth of a claim (although the whys of a claim are frequently harder to explain). Yet the main stream media still engage in pitting experts against each other or against their possible motivations. No data graphics accompanied the breast cancer screening recommendation news coverage, yet almost immediately both the news media and government officials started questioning the financial motivation of the report and the researchers. You never see this sort of behavior in a true scientific debate. Researchers do not ignore the test results from a fellow researcher because he graduated from a less prestigious school; they argue the data. This is because they both see and understand the data. Sometimes fellow researchers point out that other data may have been overlooked (and that may indeed have been accidentally due to how the primary researcher was funded) but for the most part scientific debates focus on the data and not the people.
Through data visualization the lay public can (and should) have access to the data. If done correctly we can see whether the data supports the argument, how it supports the argument and where the support may be weak. Picking apart the motivations of an expert is a crutch, a proxy for the ability to examine and question the data. In politics that method works well, but in scientific debates its a red herring. If we saw the data we could rely less on this crutch and raise the debating of the key issues of our day to a more civilized level. That’s right. I believe that news oriented data visualization can lead toward my third and final wish (an old standby):
Peace on Earth, Good Will Toward Men
Happy New Year.