Monday, September 26, 2016

Manhattan Metric

Urbica is a design firm specializing in urban data analysis. They have developed a program called Galton that graphs, for a few select cities, how far you could walk in 10 minutes (in dark blue) or 20 minutes (in lighter blue). The map of Manhattan above shows those regions for a walk originating at Broadway and 42st Street. As you walk NYC you are, for the most part, constrained to travel the grid of avenues and streets. Of course, you cannot travel as the crow flies. If you could, these regions would be concentric circles with a perimeter an equal (Euclidean) distance from your start. But walking the streets, your distance is measured by the city block metric (also known as the taxicab metric or more appropriate here the Manhattan metric). This measures distances constrained along perpendicular avenues and streets. Plotting points of equal distance with this metric from would result in the roughly rectangular (or diamond-shaped) regions shown above. Since the streets and avenues are not equally spaced and obstacles can block our travel, we don't see perfect square or rectangular regions. By the Manhattan metric, circles become squares.
Via Maps Mania.

Next week, we will see directly the results of minimizing the distance traveled in such constrained walking.

Monday, September 19, 2016

Wow, It's Hot

From xkcd, a timeline of the earth's temperature from 20,000 BCE to the present, best seen here.
When people say "the climate has changed before" these are the kind of changes they're talking about.
Since the last ice age glaciation, at no time has the climate changed faster than in the last few generations!

Monday, September 12, 2016

Maybe This Information is Beautiful, It's Just Not Accurate

This is an image from the e-book Information is Beautiful from the site of the same name. But they've made the same mistake that has long served to help people learn How to Lie with Statistics by Darrell Huff.  The graphic below, from Huff's book, shows the comparison of bags of money representing an average weekly wages from two countries. The one on the right earning twice as much as the one on the left. Showing how to lie, the bags are drawn so that the on the right is twice as tall as the one on the left, but this doesn't give us the correct visual image, since the graphic artist has doubled both the height and the width to produce the image on the right. The resulting visual impression is that the bag on the right is four times the one on the right. A perfect image to mislead.

Now the graphic from Information is Beautiful purports to represent the percentage of children in poverty. On the left is shown a small shadow silhouette of a young child with arms raised that represents 2%, the percentage of children in poverty in Denmark. Compare this with the larger silhouette for Germany representing 10%. Even in this flat 2-d outline more than 5 of the Denmark outlines could fit in the Germany outline.  The problem, of course, is that a graphic artist has lied again and doubled both the height and width of the silhouettes to represent these numbers. This distorts any comparisons that could be made with these data. And it gets worse, since we are to understand these are images of 3-dimensional children!

I recall decades ago this, now classic graphic, when it appeared in the Washington Post. The same mistake of displaying data using the same error with the same methods (compare the Eisenhower dollar with the Carter dollar). You would think....

Tuesday, September 6, 2016

Location is Time or Money

Using data from over 175,000 rentals from the real estate listing service StreetEasy, the site FiveThirtyEight asks the question:
How much would you be willing to pay to shave a minute off your commute? For New Yorkers, the answer appears to be around $56 per month. That’s how much more New Yorkers pay in rent, on average, for a one-bedroom apartment that’s a minute closer by subway to Manhattan’s main business districts.
They plot median monthly rental versus commuting time to the 42nd Street and Chambers Street subway stations. They fit, what appear to be non-parametric regressions, curves to four different groups of rentals: studios, and 1- ,2- , and 3-bedroom apartments. As expected, rental prices fall with a longer commute. From the 1-bedroom curve they estimate that, on average, a one minute shorter commute costs around $56 more a month.

Monday, August 29, 2016

Sunday, Sunday

Along with reading the Sunday newspaper, CBS Sunday Morning is a favorite in our household. Click on the video above for a short report on a streaming pay TV customer satisfaction survey conducted by JD Power. The report mentions four categories: "cord cutters", those who have cancelled TV service, "cord nevers", those who have never subscribed to pay TV and only subscribe to a streaming video service, "cord shavers", those who still subscribe but now to a downgraded TV service, and "cord stackers", those who keep pay TV but also use streaming.  From the report:
The inaugural study measures overall satisfaction among customers who have used a subscription- or transaction-based streaming video service within the past six months. The study measures customer satisfaction by examining six key measures (listed in order of importance): performance and reliability; content; cost of service; ease of use; communication; and customer service. Scores for each measure are reflected in an index based on a 1,000-point scale.
The measures for "cutters", "nevers", "shavers", and "stackers" are 802, 807, 822, and 826. Here is a frame from late in the video that reports on the right most bar "stackers" as it relates to the other categories."

It's clear that that Sunday Morning is in sore need of knowledgeable graphics editors. Perhaps their artists make their charts telegenic by filling them up with what we would call "chart junk", but in ignoring the proper representation of the data, they are presenting false and misleading images. The numbers are said to be represented here, are again, 802, 807, 822, and 826. The second bar from left almost looks twice as tall as the left-most bar and not a bar that should only be about 0.6% taller. The third bar from the left looks about 3 times taller than the left-most bar and it should only be about 2.5% taller. Finally, the right-most "cord stackers" bar should be under 3% taller, not over 4 times taller! Yes, they give a grid background to judge the sizes, but they are misleading sizes to judge. The 'smiley' satisfaction faces are probably the most reliable description of the data: The satisfaction scores are amazingly the same. From smallest to largest they vary by less than 3%.

Here are the data shown on the graph and then a more accurate rendering (without the chart junk).

Not much difference in satisfaction across all type customers.

Monday, August 22, 2016

The Wear of Love

As part of the Virginia Tourism Corporation's promotion campaign, this LOVE artwork has been placed in the Robert Reed waterfront park in Chincoteague, Virginia. It displays four, 10 foot tall Adirondack chairs. The wooden chairs spell out the LOVE with the symbols 
L O  E.
Of course, when posing for pictures, tourists prefer to sit in the      chair over the others as evidenced by the greater frequency of wear on the chairs as people climb up and rub off the paint.

This view looks at the letters in reverse. The nearest chair is E, the next is     with the most wear, then O and L. Note that the pattern of wear is in a bell-shape, with more wear near the middle of the seat and less towards the edges.

Monday, August 15, 2016

Alien Corn

Another Eastern Shore vacation outlier find. This corn stalk, amid the soybeans, is a triple threat outlier: by type, by height, and by location. Enjoy the rest of your summer.