I have huge (about 100 000) set of time series. I need to show between 5 to 10 time series, chosen semi-randomly on one chart. Chart estate is very limited - plot for each time series is only 100px x 800px.

It is not possible o predict which time series will be shown together - they have to be plot independently.

Ranges of all variables are very different: for one time series it can be between 0-3000 and for another 0 to 50. General range of all variables is 0-60000.

I want to use horizon graphs: http://timelyportfolio.blogspot.com/2012/08/more-on-horizon-charts.html but, if I make plots using 4 bands almost all time series will look flat. If I plot using 32 bands - plots will change from horizon charts, to heatmap charts.

Is it good idea, to set bands range not uniform? i.e. 1st band: 0-100, 2nd: 100-1000 3rd: 1000-5000, 4th: 5000-60000.


closed as unclear what you're asking by mkt - Reinstate Monica, kjetil b halvorsen, Michael R. Chernick, Reinstate Monica, StatsStudent Jun 21 at 4:59

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ Good question - including from a stats/visualisations view, not just programming - but I'll be interested if anyone can help - it looks like a workable solution will be very much dependent on your actual data and getting it to work within the chart estate. $\endgroup$ – Peter Ellis Feb 12 '13 at 10:05
  • $\begingroup$ I've never used, nor encountered, horizon graphs in a "real" environment, so I don't have much experience to draw from. That said, if you can't find a uniform scale for the banding, I would try a log-scale for the band values, if for no other reason than the varied scales will match what the viewers have likely encountered in other charts giving them a better frame of reference for reading your chart. $\endgroup$ – dav Mar 27 '13 at 19:17