When passing integer values to pyplot.plot, they get turned into floats, resulting in incorrect axis labels (at least, for the x-axis). For example, if you pass

x = np.arange(2006, 2012, dtype=np.int64)

as the x-value in plot, the resulting ticks are [0, 1, 2, 3, 4, 5]. I discovered this using Pandas, but was shown that it is a Matplotlib issue. Here is the original report.

pandas-dev/pandas#6272