Is there any explanation why some codepoints are narrower, and some are bellow the baseline? That seems quite weird.

Earlier this year, I added basic histograms / sparklines using these characters to a log summary script we use for quick and dirty bug triage at Wikimedia. It’s proven to be a really helpful display pattern, and I’ll probably use it elsewhere.

Dredging up some adjacent links from my browser history & bookmarks in case they’re useful for anyone else:

• https://phabricator.wikimedia.org/P14472 - an experiment with a sparklines.pl
• https://rosettacode.org/wiki/Sparkline_in_unicode
• https://metacpan.org/pod/Text::HistogramChart
• https://en.wikipedia.org/wiki/Histogram
• https://en.wikipedia.org/wiki/Geometric_Shapes