Click here to see the SAS code.
Click here to see the example.


This graph was an entry for the ...
Business Intelligence Network 2006 Data Visualization Competition

The contest was organized & judged by Stephen Few (he was the 
organizer & judge of the DM Review contest last year, which 
SAS won the dashboard competition).

This particular graph is for Question/Scenario #3 ...


At first glance, this dashboard might appear to have "too much" color.
But, rest assured, the color has a purpose! :)

Each of the plots in this dashboard has markers/bars representing
the 5 neighborhoods - I thought it was very important to be able to
identify a given neighborhood, and be able to clearly see its values
in all the graphs.  In an interactive system, I could let you click on
a neighborhood and temporarily "hilight" that neighborhood's markers
in all the plots, but in a single/static gif image, I had to color
code each neighborhood.  I picked colors which are easy to distinguish,
and which give no sense of bad/good (ie, no gradient shading).


The 3 most important things in real estate ... Location, Location, Location ;)
Why should a real estate dashboard be any different? :)

I'm not a real estate agent, but I tried to make an educated guess
that maybe "Sales Amount" and "Days on Market" might be the two
most important things, and I put them in the top row.
(A %-based sales commission is based on selling price, and 
the shorter time the houses are on the market, the more houses
an agent can sell in a year - ie, these are the factors that
most strongly affect an agent's take-home $$).

I then tried to line up charts directly below them that had the same
horizontal axes, so you could easily compare the data between the
charts by looking up/down between them.

And finally the charts along the bottom row basically take the 
info that's in the charts generally above them, and plot them 
in a slightly different way.  The visual flow should be 
generally top-to-bottom, left-to-right.


The colors are consistent in all charts, visually tying them together.
Since all the graphs use the same colors, I only need one color
legend for the entire graph.

Similar axes are lined up, when possible, to give visual continuity
and allow easy comparisons.  Sparing use of light-colored reflines
also helps in comparing between graphs (as well as seeing exactly
where the markers/bars fall on the graph).

Also, the 3 charts that have %-based axes all have a refline at 100%.
This helps show that you're looking at the same info, in different ways.


There's a "Help" button in the top/right.  In the html/web version of
this chart, clicking on that will show you a page that explains how
to read the chart.  (Information such as that the "triangle" marker
shows last year's median value.)  This explanatory, which users 
generally only need to see once, would otherwise clutter the chart.

Although the contest is just for the image, this chart was also created
with html charttips (hover your mouse over the colored markers & bars
and you'll see the data values), and drilldowns (click on the markers
and bars, and it drills-down and shows you the spreadsheet of the 
actual/raw data).


Results / Observations:

Filthy Richlands has the highest Sale Amount, and its median
has risen dramatically since last year.  (Likewise for the
Asking Price).

Badlands Sale Amount is frequently higher than the Asking Price
(so much so that the median is actually over 100%).

Badlands has the lowest median days on market, and 
Somnolent Community has the highest median days on market.


Back to Samples Index