Our new provider of heating gas in Colorado is Xcel Energy. They have an elaborate Flash-based website for customers which includes this graph of usage history:
Xcel Usage Graph
I see three key problems here.
First, the “3D” effect is widely known to distort perceptions of data. It’s too easy, for example, to see those last two dark bars as depicting similar magnitudes since the leading edge of one is so close to the trailing edge of the other. This chart has a single dimension of data (“cost”), so it needs a one-dimensional presentation. “Pretty” is confusing — what Edward Tufte calls “chartjunk.”
Second, the chart overlaps data from three years, emphasizing comparisons between costs in the same month of different years, while making the long-term trend difficult to identify. Are we actually spending less this year than last? This arrangement of data is particularly troubling since peak usage is naturally in the winter, which always spans two calendar years. Having so much historical data is a privilege; let’s see it all laid out chronologically.
Finally, what’s plotted is not actually “usage” at all, but “cost.” It’s impossible to tell from this chart alone whether we used less energy this year or if rates fell. In fact, our dramatically higher costs in 2008 were from a previous residence where Xcel provided both gas and electricity. The cost was higher, even if our gas usage was lower. We need charts that show our energy usage (in kilowatt hours or in cubic meters, as appropriate), not how much money we’ve spent.
I still like when airline pilots turn on the public address system and tell us about the flight. A few even still announce interesting landmarks (“passengers on the left side of the plane will have a great view of New York City”), but most limit themselves to a brief “welcome aboard” message and another quick warning close to landing (“I’ll be turning on the ‘Fasten Seat Belt’ sign in a few minutes…”).
I don’t need constant narration, but I enjoy the faint reminder of an era when jet travel was glamorous and passengers in sandals were unthinkable. Having said that, however, pilots need to think carefully before giving weather reports. It’s not a bad gesture, but they usually get a bit carried away.
A frustrating percentage of pre-landing weather announcements sound like this:
The temperature on the ground in Boston is 37° Fahrenheit, with some cloud cover, winds from the southwest at 10 miles per hour, and five miles visibility.
Five miles visibility.
Honestly, I’m glad the pilot is aware of that — it seems like useful information to have handy when flying an airplane — but is there anybody else on board who cares? Even other pilots who happen to be in the cabin really don’t need to know at that moment.
I bring this up not because it happened once on a flight (which would have been amusing, though not particularly memorable) but rather because it seems to happen on every flight. Respectable and talented pilots have told countless people across the country how far in front of them they could see, and it’s time for the practice to end.