istats - bash script to get various statistics from numerical data on stdin

istats

#!/bin/bash
# iStats - get various statistics from numerical data on stdin
# aka "incredibly easy Stats" aka. "Stats for idiots" ;)
# fork of "jstats2" https://gist.github.com/roblogic/da238766f5e645ddfb7290b06b1f80f0
set -e     # Exit immediately on error
# set -x   # Debug
usage="<input-stream> | istats"

# Create temporary file
WORK=$(mktemp /tmp/$(basename $0).XXXXX) || exit 1
sort -n $* > $WORK

# Count data points
N=$(wc -l < $WORK)
[ $N -eq 0 ] && { echo "$usage" ; exit 1 ; }
echo " --stats analysis--"
echo " count n  $N"

# Find percentile locations
P50=$(dc -e "$N 2 / p")
P90=$(dc -e "$N 9 * 10 / p")
P95=$(dc -e "$N 95 * 100 / p")
P99=$(dc -e "$N 99 * 100 / p")

# Print percentile values etc
echo "  median  $(awk "FNR==$P50" $WORK)"
echo "     p90  $(awk "FNR==$P90" $WORK)"
echo "     p95  $(awk "FNR==$P95" $WORK)"
echo "     p99  $(awk "FNR==$P99" $WORK)"
echo -e "  mean \u03BC  $(awk '{sum+=$1}END{if (NR>0) print sum/NR}' $WORK)"
echo -e "stddev \u03C3  $(awk '{x+=$0;y+=$0^2}END{print sqrt(y/NR-(x/NR)^2)}' $WORK)"
echo "     min  $(awk "FNR==1" $WORK)"
echo "     max  $(awk "END{print}" $WORK)"
echo "datafile  $WORK"

Usage Example

$ seq 999 | istats
 --stats analysis--
 count n  999
  median  499
     p90  899
     p95  949
     p99  989
  mean μ  500
stddev σ  288.386
     min  1
     max  999
datafile  /tmp/istats.E4Gd0