03.12.2017

The Advent of Void: Day 3: ministat

On day 3 we introduce you to the small statistics tool ministat(1) that originates from FreeBSD.

When we sling around data on the command line, sometimes we want to compute some basic statistics on it. While minimum, average and maximum are easily computed with awk(1), you probably forgot the formula for more complicated statistics. There is no need to pull out R already, lets use ministat instead!

For example, let’s have a look at the historical summer temperatures in Germany. We need to strip off the header, and then we can look e.g. at the measurements in Bavaria:

% sed 1,2d <regional_averages_tm_summer.txt | ministat -d';' -C6
x <stdin>
+------------------------------------------------------------------------------+
|                    x       x                                                 |
|                    x       x                                                 |
|                x   x       x      x                                          |
|                x   x       xx     x                                          |
|              x x   x       xx  x  x                                          |
|              x x   xx   x  xxx xx x                                          |
|              xxx   xx x x xxxx xx x  x     x                                 |
|          x  xxxx xxxx x x xxxx xxxx xx   x x                                 |
|        x x  xxxx xxxx xxx xxxx xxxx xxx  x x  x      x                       |
|x     x x xx xxxx xxxx xxx xxxx xxxx xxxx x xx xx x x x  x      x            x|
|                |___________A___________|                                     |
+------------------------------------------------------------------------------+
    N           Min           Max        Median           Avg        Stddev
x 137          13.9            20          16.1     16.105839    0.96883545

ministat will prepare a nice ASCII-art plot so we can get a feel for the data distribution. We have 137 measurements, and an average temperature of 16.1°C.

To compare multiple datasets, we need to pass them as multiple input files. Luckily, our shell has process substitution. Let’s match the Bavarian measurements with the German average:

% ministat -s <(awk -F';' 'NR>2{print $6}' regional_averages_tm_summer.txt) \
              <(awk -F';' 'NR>2{print $19}' regional_averages_tm_summer.txt) 
x /proc/self/fd/11
+ /proc/self/fd/12
+------------------------------------------------------------------------------+
|                    x       x        +                                        |
|                    x       x        +                                        |
|                x   x      +x   +  x +                                        |
|                x   x   +  +x*  + +x +                                        |
|              x x   x   +  +x*  *++* +  +                                     |
|              x+x   *x  +x +**x **+* ++ +      +                              |
|              x*x + *x x+x **** **+* +* + + x  +                              |
|          x  xx*x *x*x *+x **** **** ** + * x  + +                            |
|        x *  *x** **** **x **** **** ***+ * *  *++    x ++                    |
|x     x x *x **** **** *** **** **** **** *+** **+* *+x +*+     x        +   x|
|                |___________A___________|                                     |
|                     |___________A__________|                                 |
+------------------------------------------------------------------------------+
    N           Min           Max        Median           Avg        Stddev
x 137          13.9            20          16.1     16.105839    0.96883545
+ 137          14.7          19.7          16.5     16.481752    0.89838245
Difference at 95.0% confidence
	0.375912 +/- 0.221251
	2.33401% +/- 1.37373%
	(Student's t, pooled s = 0.934273)

ministat now plotted both datasets, and computed a Student’s t-test which shows it’s indeed a bit colder in Bavaria than in Germany in general.

Of course, in winter its even colder.