plot

A simple but useful wrapper around some basic pyplot functionality.

Reads data from stdin and pops up a window showing a plot of it. It's designed to be easy to use with the sort of data that can be munged into the appropriate format with sed, awk, cut, etc. Data elements can be whitespace- or comma-separated.

Plot types

plot TYPE can generate the following TYPEs of graphs:

• Line graphs (line): expects (x,y) pairs or single y-values

• Bar graphs (bar): expects (label,value) pairs

• Histograms (hist): expects single values

• Scatter plot (scatter): expects (x,y) pairs

• Cumulative distribution (cdf): expects single values

• Timechart (tc): expects (label,start,end) or (label,start,length) tuples

• Heatmaps (heatmap): expects equal-length multi-value lines forming a 2D array

• Violin plots (violin): expects multi-value lines

Examples

To show a 20-binned histogram of git commit frequency:

$ git log --date=short --pretty=tformat:"%ad" | uniq -c | plot hist --bins 20

Scatter plot of line count vs. file size:

$ wc -lc * | head -n-1 | plot -x Lines -y Bytes scatter

Cumulative distribution of word length

$ awk '{ for (i=1; i<=NF; i++) print length($i) }' wrnpc.txt | plot cdf -b20

TODO

• pyplot kwargs in command-line arguments?

• Generate-a-script mode? (i.e. output python code for further editing/reuse)

Author

Zev Weiss zev@bewilderbeest.net