One of the effective way to present a time series data over long period of time is either typical line charts or some modified versions of it. It gets little harder to visualize when would like to do see clustering of data points. Here is one where I find it tiled series that gives quick glimpse of a some metric varying over few years. This is a daily metric starting January 01, 2010 and up to recent week of May, 2012; nearly 2 and 1/2 years of metric variation.
The Y-axis measures represent day of the week (wday) with 1 representing Monday, 2 Tuesday and so forth with 7 as Sunday. I set the values to these so that I could cluster weekends metrics (5,6 & 7) together vs the weekday (1,2,3,4 - Monday to Thursday). X-axis represent week of the year (1 to 52 or 53). Year 2010, 2011 and 2012 are series. Metric values varies from 0.00 to 10.00 and color of each tile varies slightly based on the metric value.
If you were to use discrete values of metrics say, 0 to 5 the color coding is similarly quite distinct. See graph 2 below. The data is from 2nd week of March, 2010 to May, 2012.
The Y-axis measures represent day of the week (wday) with 1 representing Monday, 2 Tuesday and so forth with 7 as Sunday. I set the values to these so that I could cluster weekends metrics (5,6 & 7) together vs the weekday (1,2,3,4 - Monday to Thursday). X-axis represent week of the year (1 to 52 or 53). Year 2010, 2011 and 2012 are series. Metric values varies from 0.00 to 10.00 and color of each tile varies slightly based on the metric value.
If you were to use discrete values of metrics say, 0 to 5 the color coding is similarly quite distinct. See graph 2 below. The data is from 2nd week of March, 2010 to May, 2012.
Graph 1
Graph 2
