Uncategorized

Seven questions a statistician could answer for a lawyer

A statistician could help a lawyer answer the following questions.

  1. Was this data collected in a proper way?
  2. Does common sense apply here, or is there something subtle going on?
  3. What conclusions can we draw from the data?
  4. Is this analysis routine or is there something unusual about it?
  5. How much confidence can we place in the conclusions?
  6. How can you explain this to a jury?
  7. Did opposing counsel analyze their data properly?

Related:

Visualizing the DFT matrix

The discrete Fourier transform (DFT) of length N multiplies a vector by a matrix whose (j, k) entry is ωjk where ω = exp(-2πi/N), with j and k running from 0 to – 1. Each element of the matrix is a rotation, so if N = 12, we can represent each element by an hour on a clock. The angle between the hour hand and minute hand corresponds to the phase of the matrix entry. We could also view each element as a color around a color wheel. The image below does both.

The matrix representing the inverse of the DFT is the conjugate of the DFT matrix (divided by Nf, but we’re only looking at phase here, so we can ignore this rescaling.) The image below displays the DFT matrix on the left and it’s inverse on the right.

Taking the conjugate amounts to making all the clocks run backward.

The DFT is often called the FFT. Strictly speaking, the FFT is an algorithm for computing the DFT. Nobody computes a DFT by multiplying by the DFT matrix, because the FFT is faster. The DFT matrix has a lot of special structure, which the FFT takes advantage of to compute the product faster than using ordinary matrix multiplication.

By the way, there are Unicode characters for clock times on the hour, U+1F550 through U+1F55B. I created the image above by writing a script that put the right characters in a table. The colors have HSL values where H is proportional to the angle and S = L =0.8.

Click to learn more about consulting help with signal processing

Finding 2016 in pi

2016 appears in π starting at the 7173rd decimal place:

You can confirm this with Mathematica or Wolfram Alpha:

        Mod[ Floor[10^7177 pi] , 10000]

I found it using the following Python code:

        >>> from sympy import pi
        >>> digits = str(pi.evalf(10000))[2:]
        >>> digits.find('2016')
        7173

By the way, it’s also true that 2016 = 1 + 2 + 3 + … + 63.

Automate to save mental energy, not time

Automation doesn’t always save as much time or effort as we expect.

The xkcd cartoon above is looking at automation as an investment. Does the work I put in now eventually save more work than I put into it? Automation may be well worth it even if the answer is “no.”

Automation can be like a battery as well as an investment. Putting energy into batteries is a bad investment; you’ll never get out as much energy as you put in. But that’s not why you put energy into batteries. You put energy in while you can so you can use some of that energy later when you need it.

Write automation scripts when you have the time, energy, and motivation to do so and when nothing else is more important. (Or nothing is more interesting, if you’re looking for a way to procrastinate without feeling too guilty. This is called “moral compensation.”) You may indeed save more work than you put into writing the scripts. But you also may save mental energy just when you need it.

Suppose it takes you an hour to write a script that only saves you two minutes later. If that two minutes would have derailed your concentration at a critical moment, but it didn’t because you had the script, writing the script may have paid for itself, even though you invested 60 minutes to save 2 minutes.

If your goal is to save mental energy, not time, you have a different strategy for automation. If a script executes faster than a manual process, but it takes a long time to remember where to find the script and how to run it, it may be a net loss. The less often you run a script, the chattier the interface should be.

The same considerations apply to learning third party software. I suspect the time I’ve put into learning some features of Emacs, for example, will not pay for itself in terms of time invested versus time saved. But I’ve invested leisure time to save time when I’m working hard, not to save keystrokes but to save mental energy for the project at hand.

Related post: How much does typing speed matter?

Click to learn more about math and computing consulting

Restarting @DSP_fact, ending @PerlRegex

I’m making a couple changes to my Twitter accounts.

First, I’m winding down @PerlRegex. I’ll stop tweeting there when my scheduled tweets run out. I suggest that everyone who has been following @PerlRegex start following @RegexTip instead. The latter is more general, but is mostly compatible with Perl.

Second, I’m reviving my @DSP_Fact. I stopped tweeting there a couple years ago, but I’d like to start posting there again. This time it’s going to be a little broader. I intend to include some material on acoustics, Fourier analysis (continuous and discrete), and maybe some other related material.

DSP_fact logo

Alternating sums of factorials

Richard Guy’s Strong Law of Small Numbers says

There aren’t enough small numbers to meet the many demands made of them.

In his article by the same name [1] Guy illustrates his law with several examples of patterns that hold for small numbers but eventually fail. One of these examples is

3! – 2! + 1! = 5

4! – 3! + 2! – 1! = 19

5! – 4! + 3! – 2! + 1! = 101

6! – 5! + 4! – 3! + 2! – 1! = 619

7! – 6! + 5! – 4! + 3! – 2! + 1! = 4421

8! – 7! + 6! – 5! + 4! – 3! + 2! – 1! = 35899

If we let f(n) be the alternating factorial sum starting with nf(n) is prime for n = 3, 4, 5, 6, 7, 8, but not for n = 9. So the alternating sums aren’t all prime. Is f(n) usually prime? f(10) is, so maybe 9 is the odd one. Let’s write a code to find out.

    from sympy import factorial, isprime

    def altfact(n):
        sign = 1
        sum = 0
        while n > 0:
            sum += sign*factorial(n)
            sign *= -1
            n -= 1
        return sum

    numprimes = 0
    for i in range(3, 1000):
        if isprime( altfact(i) ):
            print(i)
            numprimes += 1
    print(numprimes)

You could speed up this code by noticing that

    altfact(n+1) = factorial(n+1) - altfact(n)

and tabulating the values of altfact. The code above corresponds directly to the math, though it takes a little while to run.

So it turns out the alternating factorial sum is only prime for 15 values less than 1000. In addition to the values of n mentioned above, the other values are 15, 19, 41, 59, 61, 105, 160, and 601.

* * *

[1] The Strong Law of Small Numbers, Richard K. Guy, The American Mathematical Monthly, Vol. 95, No. 8 (Oct., 1988), pp. 697-712.

For daily tips on Python and scientific computing, follow @SciPyTip on Twitter.

Scipytip twitter icon

Non-technical books I’ve written about this year

Here are some of the non-technical books I’ve mentioned in blog posts this year. I posted the technical list a couple days ago.

Maybe I should say “less technical” rather than “non-technical.” For example, Surely You’re Joking, Mr. Feynman is a book about a physicist, but it’s at least as much a human interest book as a science book.

 

 

Four ways to find hidden RSS feeds

RSS feeds

RSS lets you subscribe to blogs. It also lets you read posts in peace, free from distracting peripheral ads. This explains why Google would kill off the world’s most popular RSS reader.

Blogs used to display an icon linking to the site’s RSS feed, and any still do. Blogging software still creates RSS feeds, though links to these feeds have become harder to find.

There are several ways to find the RSS feed of a site that does not make this feed obvious.

  1. Your blog reader may be able to find the RSS from the site’s URL.
  2. You can try adding /feed or /rss to the end of the domain.
  3. You may be able to find the RSS feed by looking at the HTML source of the front page.
  4. There are browser plug-ins that will show when a page has RSS feeds.

Example

Take for example Bill Gates’ blog. You’ll see at the top how to subscribe by email, and you’ll see at the bottom links to Facebook, Twitter, LinkedIn, and YouTube, but no RSS.

Digg Reader was able to find the RSS feed from just the top level URL http://www.gatesnotes.com/.

Tacking /feed on the end doesn’t work, but http://www.gatesnotes.com/rss does.

If you open up the page source, you’ll find 31 references to RSS. Turns out there is a link to the RSS feed after all. It’s in a menu under the + sign in the top right corner. While writing this, I started to change my example since I was wrong about the site not displaying an RSS link. But I decided to keep it because it shows that the steps above also work when there is an RSS link but you’ve overlooked it.

My RSS feeds

If you want to subscribe to this blog via RSS, there’s a big blue button on the right side that says Subscribe by RSS. You can also subscribe to individual categories of posts in case you’d like to subscribe to my math posts,  for example, but not my software development posts, or vise versa.

Subscribe to blog by RSS

There are also RSS feeds to my Twitter accounts, thanks to BazQux.

 

Technical books I’ve written about this year

Here are some of the books I’ve mentioned in blog posts this year. One of these books may be just the present for a geek in your life.

This post looks at technical books: math, science, engineering, programming. I’ll have a follow-up post with non-technical books I’ve written about. (Update: here’s the non-technical list.)

Programming

Science and engineering

Math

 

Colors of noise

visible spectrum

The term white noise is fairly common. People unfamiliar with its technical meaning will describe some sort of background noise, like a fan, as white noise. Less common are terms like pink noise, red noise, etc.

The colors of noise are defined various ways, but they’re all based on an analogy between the power spectrum of the noisy signal and the spectrum of visible light. This post gives the motivations and intuitive definitions. I may give rigorous definitions in some future post.

White noise has a flat power spectrum, analogous to white light containing all other colors (frequencies) of light.

Pink noise has a power spectrum inversely proportional to its frequency f (or in some definitions, inversely proportional to fα for some exponent α near 1). Visible light with such a spectrum appears pink because there is more power toward the low (red) end of the spectrum, but a substantial amount of power at higher frequencies since the power drops off slowly.

The spectrum of red noise is more heavily weighted toward low frequencies, dropping off like  1/f2, analogous to light with more red and less white. Confusingly, red noise is also called Brown noise, not after the color brown but after the person Robert Brown, discoverer of Brownian motion.

Blue noise is the opposite of red, with power increasing in proportion to frequency, analogous to light with more power toward the high (blue) frequencies.

Grey noise is a sort of psychologically white noise. Instead of all frequencies having equal power, all frequencies have equal perceived power, with lower actual power in the middle and higher actual power on the high and low end.

Click to learn more about consulting help with signal processing

Twitter account wordclouds

Here are wordclouds for some of my most popular Twitter accounts. Thanks to Mike Croucher for creating these images. He explains on his blog how to create your own Twitter wordclouds using R.

My most popular account is CompSciFact, tweets about computer science and related topics.

AlgebraFact is for algebra, number theory, and miscellaneous pure math. (Miscellaneous applied math is more likely to end up on AnalysisFact.)

ProbFact is for probability.

DataSciFact is for data science: statistics, machine learning, visualization, etc.

You can find a full list of my various Twitter accounts here.

Mathematical alchemy and wrestling

David Mumford wrote a blog post a few weeks ago in which he identified four tribes of mathematicians. Here’s a summary of his description of the four tribes.

  • Explorers are people who ask — are there objects with such and such properties and if so, how many? …
  • Alchemists … are those whose greatest excitement comes from finding connections between two areas of math that no one had previously seen as having anything to do with each other.
  • Wrestlers … thrive not on equalities between numbers but on inequalities, what quantity can be estimated or bounded by what other quantity, and on asymptotic estimates of size or rate of growth. This tribe consists chiefly of analysts …
  • Detectives … doggedly pursue the most difficult, deep questions, seeking clues here and there, sure there is a trail somewhere, often searching for years or decades. …

I’m some combination of alchemist and wrestler. I suppose most applied mathematicians are. Applications usually require taking ideas developed in one context and using them in another. They take often complex things then estimate and bound them by things easier to understand.

One of my favorite proofs is Bernstein’s proof of the Weierstrass approximation theorem. It appeals to both alchemists and wrestlers. It takes an inequality from probability and uses it in an entirely different context, one with no randomness in sight, and uses it to explicitly construct an approximation satisfying the theorem.

I thought of David Mumford’s tribes when I got an email a couple days ago from someone who wrote to tell me he found in one of my tech reports a function that he’d studied in his own research. My tech report was motivated by a problem in biostatistics, while he was looking at material structural fatigue. The connection between remote fields was a bit of alchemy, while the content of the tech report, an upper bound on an integral, was a bit of wrestling.