Posts tagged ‘Flowing Data’

Solar False Advertising

I saw this at Flowing Data -- this is apparently a chart prepared by some sustainability group at MIT to map solar potential of different sites in Cambridge, MA

Look at all the sites marked "excellent".  I have news for the brilliant folks at MIT.  Even the best, flattest roof facing south in Cambridge, MA still rates a "sucks" for solar potential. (source)

Even with massive state and Federal subsidies, those of us who live in the bright red areas find that roof-top solar PV is still an - at best - marginal investment with very long payback times.  We all hope to change this in the future, but there is no way a city like Cambridge with approximately half the solar insolation we get in AZ is going to have "excellent" roof top solar PV sites.

State Stereotypes

This is pretty awesome.  Using Google's auto-suggest which is based on their most frequent searches, Renee DiResta created a rollover map of state stereotypes.  Here is Arizona's, the rest are here.  Via Flowing Data

Reservoir of Meyers

I thought this map in the National Geographic was cool- it shows the most popular surname in different areas of the country.  I am not sure what geographic divisions they use (why does Texas have so little granularity?).  But it does turn out there are a couple of "Meyer" labels on the chart, which kind of surprised me  -- though they did spell it the right damn way - no "s" on the end, no Mayer, no Myer or other such nonsense ;=)

The "Meyer" in the middle of Iowa is pretty much where my dad's family settled when they came over from Northern Germany.  Though I am confused as to why they show it color-coded as English origin -- I am pretty sure this area is German  (the Meyer over around northern Nebraska and in Wisconsin is coded German).

HT:  Flowing Data

Stock Market Returns

This chart in the NY Times is pretty interesting, though I could quibble about the color coding.  You have to stare at it a minute to get it - each cell represents a combination of stock purchase and sales dates, with the color representing the average market inflation-adjusted return for that buy and hold period (click to enlarge, or click through to the source link where it is explained in more depth).

Whenever one uses red and green for coloring a chart, the reader is going to assume red is bad and green is good.  In this case, the light red represents returns from 0 to 3% above inflation.  Is that bad?  Maybe.  I would say inflation plus 3% is probably lower than people's expectation of stock market returns, but I think a lot of folks would equate red with capital erosion, which is not the case if returns are out-pacing inflation.

This is sort of a good-news-bad-news story.  The good news is that there is no 25-year period where returns fall below inflation.  The bad news is that the median return of inflation plus 4% is probably less than most folks are planning for -- including a lot of state pension funds that are still counting on returns like 8% for their entire portfolio (something like inflation + 5-6%), which is a blend of stocks and bonds, implying they are hoping for an equity return north of that.

HT:  Flowing Data

Changing Flow of the Mississippi River

I thought this was an incredibly cool image, showing the changing path of the Mississippi River (in this case where it meets the Ohio).  (via Flowing Data)

When I was a kid, I was fascinated by water flow and erosion.  I remember spending a whole day on a woodside hill watching the evolution of an ad hoc stream of water, playing around with damming it in some places, creating new channels, etc.  When I went to the beach, I never built castles but attempted to build walls and channels to shape the way the tide flowed.  Since I am free associating, I also remember visiting a huge model of the Mississippi, I think near Vicksburg, that I thought at the time was the coolest thing on Earth.  Not even sure today if it still exists.

Stuff [Race][Gender] Like

This is pretty interesting -- OK Cupid did a phrase-frequency count in its online dating ads and were able to sort them by race and gender, and then identify those phrases that the particular race-gender combination used most uniquely.  Its kind of amazing just how much the analysis might fit your stereotyped guesses.  Among many others, horseback riding, baths, and Jodi Picoult for white women, with Tom Clancy, Harleys and Soundgarden for white men.  Check it out here.  (hat tip Flowing Data)

Some Geek Love

One of the geekier conversations I used to get drawn into (up there along with arguing about favorite Serenity episodes, lamenting the demise of Omni magazine, and coming to blows over D&D rules interpretations) was over the relative merits of various sorting algorithms.  Flowing Data has some links to some interesting visualization approaches to sorting algorithms.  This one is for quicksort (colors start out random on the left and must be sorted into Roy G Biv order).

In college I did a project on solving traveling-salesman-type problems with an algorithm called simulated annealing.  Many approaches to the traveling salesman problem pick a random path and then randomly switch pairs of routings on the path, and then stick with the alternative that gives the best score (in this case the shortest path).   Rinse, repeat a zillion times.  The problem is this approach can get stuck in, or converge on, a local optima that are not as good as the single grand optimum for the problem.

In simulated annealing, the algorithm is allowed to sometimes jump to a worse (ie longer) path, which lets it jump out of local minima.  The amount of the backwards jump that is allowed is slowly reduced over time as the algorithm runs.  It is called simulated annealing because this is very similar to the annealing process in metals, where temperature is decreased slowly to, initially, allow the metal molecules to jump to higher energy states so that the whole can settle into a more homogeneous structure.

Anyway, we tried to show how the algorithm proceeded to a solution over time and the visualizations looked a little like this.

Infographic of the Day

It's either this one, via Flowing Data

Can't you picture some Federal bureaucrat with purview somehow over wood pallet fires trying to fan the flames of public opinion in the interests of his or her job security and budget?

The other candidate is this from an outstanding XKCD post on a color naming survey he did (via TJIC).

The whole post is hilarious and worth you time.

Why Modern Music Sucks

Boy, do I sound like my parents with that headline, or what?

Apparently, two kinds of compression are changing the sound of recorded music.  The first is digital compression, such as we use to get a bunch of mp3's on an iPod.    I still buy CD's, and then rip them myself so I can control the bit rate and compression, but a lot of folks are buying mp3's online of all kinds of quality.  (I actually rip every CD twice -- once as a VBR MP3 for my iPod and once as a loss-less FLAC file for my home audio server).

The second type of compression, perhaps more insidious because it is impossible for the individual listener to control, is use of audio compressors that reduce the dynamic range of music - basically making soft parts louder and vice versa.    NPR discusses it here, via Flowing Data. While the second form of compression is as old as vinyl (the revenge of Phil Specter?) these two types of compression are related as apparently louder music gives more room to hide digital compression artifacts, so producers are compressing music and increasing loudness.

The best test I have of dynamic range is listening to music in a noisy car, say with the windows open.  Many classical disks can't be listened to this way, as the variation from soft to loud causes one to keep having to fiddle with the volume knob.  I have a few old rock disks that have the same kind of range  (some old Genesis albums come to mind) but most of my newer disks will play just fine in a loud car, probably meaning that they indeed do have much narrower dynamic ranges.

To some extent, this is counter intuitive to me given the prevalence of headphone listening, since headphones are great for listening to music with big dynamic ranges.  But what do I know?  I grew up listening to 8-tracks so it all is an improvement for me.

Here is a very good, succinct example of how compression works and why it makes music suck:

Mariano Rivera in the Playoffs

Via Flowing Data, a cool chart in the NY Times with every batter faced by Rivera in the playoffs.

Repeating the Same Mistake, Over and Over

Flowing Data draws my attention to this nutty chart in the New Scientist  (I have never read the New Scientist, but my experience is that in periodicals one can generally substitute "Socialist" for the word "New").  Click to enlarge.


Will the world really run out of Indium in 5 years?  Of course not.  New sources will be found.  If they are not, then prices will rise and a) demand with be reduced and b) efforts to find new sources will be redoubled.  Push come to shove, as prices rise too much, substitutes will be found (which is why John D. Rockefeller probably saved the whales).  Uranium is a great example -- sure, proved reserves are low right now, but companies that mine the stuff know that there is tons out there.  That is why they are going out of business, there is too much supply for the demand.  Any spike in price would immediately generate tons of new developed resources.  And even if we run out, there are enormous quantities of thorium which is a potential substitute in reactors.

Absolutely no one who was old enough to be paying attention to the news in the 1970s could have missed charts very similar to this.  I remember very clearly mainstream articles that we would run out of oil, titanium, tungsten, etc. by the early 1990's.  Seriously, name one commodity we have plain run out of (*cough* Julian Simon *cough*).

People say, well, the resources have to be finite and I would answer, "I suppose, but given that we have explored and mined about 0.000001% of the Earth's crust and none of the floating mineral reservoirs in space (called asteroids), I think we are a long, long way from running out."

You would think that the guys running this analysis would get tired of being so wrong so consistently for so many decades, but in fact their real point is not about resources but about the US and capitalism.  The point of the chart is not really to say that the world will credibly run out of tungsten, but to tell the world that it is time to get out their pitchforks because the US is stealing all their wealth and resources.  It is an age-old zero-sum wealth fallacy that has never held any water, but remains a powerful talking point among socialists none-the-less.

For socialists, wealth is not created by man's mind and his effort -- it is a spring in the desert with a fixed flow rate.  It just exists to be taken or fought over.  The wealthy, by this theory, have not earned their wealth, they are just the piggy ones who crowd to the front of the line and take more than their share from the spring.  Unfortunately, socialists have never been able to explain why the spring, which flowed so constantly (and so slowly) for thousands of years, suddenly burst forth with a veritable torrent in lockstep with the growth of capitalism in the west.  And why it seems to dry up in countries that adopt socialism.

Postscript: A while back I posted on the New Economics Foundation  (remember what I said about "New") and their claim the world had just gone into ecological debt.