Our friend Marylaine Block once again delivers some insights directly applicable to analytic journalism.  See the piece below where she explains why visual statistics and infographics are essential to what we're doing (or trying to do).



ExLibris #268  Permanent URL http://marylaine.com/exlibris/xlib268.html
Archive: http://marylaine.com/exlibris/archive.html

THE POWER OF VISUALIZED INFORMATION
by Marylaine Block

When I discussed some possible futures for reference service at the
California Library Association <http://marylaine.com/ref.html>, I focused
heavily on the value we create for users by not just finding information
for them but providing context and meaning for information. One of the best
ways to do this is by presenting it visually.

This is especially important when we're talking about numbers, because the
human mind is poorly equipped to grasp the meaning of large numbers. Any
number higher than those we have worked with in our personal lives, like
the amount of our salary or our mortgage, are, for all intents and
purposes, classified together in our minds as "a whole bunch." The real
meaning of millions, billions, and trillions is effectively beyond our
grasp (and maybe beyond the grasp of legislators who routinely deal in
these numbers); That's why I like to point people to the Megapenny project,
<http://www.kokogiak.com/megapenny/default.asp>, which visually demonstrates
the substantial difference between million billion, and trillion.

Numbers conveyed in charts are more readily graspable and have more
dramatic impact than row after row of numbers in eye-glazing tables.
Consider the nice charts OCLC has provided for librarians to demonstrate
the economic impact of libraries,
<http://www5.oclc.org/downloads/community/librariesstackup.pdf>. The visual
demonstration of how visits to libraries exceed attendance at all
professional and collegiate sports by a factor of five is a splendid
response to the question, "With Google, who needs libraries anymore?"

Take a look at how somebody displayed the results from mining data about
political books from "readers who bought this also bought these" systems at
major web booksellers: <http://www.orgnet.com/divided.html>. That graphic
representation powerfully conveys the findings in a few seconds; the
details can be read at your leisure.

Consider also how librarians at Cornell University's Engineering Library
explained to their faculty the problem of excessive and escalating sci-tech
journal prices, <http://www.englib.cornell.edu/exhibits/stickershock/>.
(Librarians, of course, are the fools publishers can count on to buy The
Journal of Applied Polymer Science rather than the Toyota Corolla.) This
visual demonstration was an important tool librarians used to convince
faculty to join the fight to control the costs of scholarly publishing.

Those of us who have frequently used reference books like The Timetables of
History, or Who Was When already understand the way that concurrent visual
timelines can contextualize any subject. Nothing exists in a vacuum, and
art, literature, music, science, and historical events coexisting at the
same time inevitably influence each other. The history of medicine and the
history of photography have seen significant advances in wartime, for
example. The music of Wagner and the philosophy of Nietzsche had a powerful
ipact on the development of the National Socialist party in Germany. To
help our users understand those coexisting influences, you can send them to
concurrent timeline sites like HyperHistory,
<http://www.hyperhistory.com/online_n2/History_n2/a.html>

Mapping is another valuable way of providing context for information. The
Perry-Castaneda Library Map Collection at the University of Texas helps
illuminate current news stories by providing current and historical maps
<http://www.lib.utexas.edu/maps/>. Consider how much more comprehensible
the conflict in Iraq is when you view maps that show the Distribution of
Ethnoreligious Groups and Major Tribes, or Land Use, or the distribution of
oil facilities <http://www.lib.utexas.edu/maps/iraq.html>.

When people need information specific to their own community, Google Maps
<http://maps.google.com/> allows you to
create localized topical maps easily. Feed in "Restaurants near
AddressOfYourLibrary" and you'll get a map you can duplicate and hand out
to your patrons (which I would urge you all to do).

As people ask you for local information, consider whether they'd benefit
from having you display it as a Google map. Here are just a few of the ways
people have been using Google Maps: to map the locations for best gas
prices (<http://www.ahding.com/cheapgas/>); public transit stops near a
given location (see <http://holovaty.com/blog/archive/2005/04/19/0216>);
traffic information (see <http://traffic.poly9.com/>); sex offenders (see
<http://www.mapsexoffenders.com/>); Wireless Hotspots (see
<http://www.tadl.org/wireless/map/>). I'm sure you can think of lots more
uses.

A particularly powerful form of mapping is Geographic Information Systems
(GIS), which the GIS Dictionary at ESRI defines as "an integrated
collection of computer software, spatial data, related information, and
supporting infrastructure used to visualize and analyze spatial
relationships, model spatial processes, and manage spatial information."
(See <http://www.gis.com/> and
<http://www.library.wisc.edu/data/GIS/gisrsrc.htm> for more information on
GIS). By allowing you to superimpose on each other multiple types of
information with geographic coordinates, it's a powerful tool for analyzing
relationships between data -- between, say, a community's geology,
drainage, and proposed development, or between a library's buildings, its
service area, and the demographic communities within it.

A necessary caveat because of the very power of graphic representations,
however, is their capability for distorting information. We knew this even
before people started using PhotoShop to alter images. After all, the mere
fact of where you choose to stand to take a picture and what you select to
shoot alters the "reality" revealed by the picture; those choices allow you
to make a demonstration sparsely attended, or so big it shut the city down,
or to make its participants everyday middle-class people, or obvious
radicals and nutcakes.

Consider the famous red state/blue state map
<http://www-personal.umich.edu/~mejn/election/>. Because this map
represents physical space occupied by states awarded under a
winner-take-all electoral system, it appears to show Democratic voters
hanging on by their fingernails to the edges of a continent that is
rejecting them.

Arrow down through that web site and you'll see that, since much of that
physical red-state space has more cows than people, a cartogram that skews
the size of the states to correspond to the population of those states
provides an entirely different view. Arrow down still further and you'll
understand how, with electoral votes awarded by state, the
red-state-blue-state depiction made states with substantial pockets of both
red and blue voters look more monolithic than they actually are; the
speckled county by county map gives a far better presentation of a country
that's not so much red and blue as a mix of both.

That's why when we use a tool as powerful as graphics to illuminate
information, it's especially incumbent on us to document and explain our
sources and methods fully, and to explain any assumptions embedded in the
data as imaged. It's our obligation, as information professionals, to honor
the data, and to honor our users.


* * * * *