This gist illustrates the issue in Plotly of non-json-serializable masked constants as produced by Pandas with and without the pull request at https://github.com/plotly/python-api/pull/159) (also see http://stackoverflow.com/questions/27053729)

pandas_masked_numbers.py

import pandas as pd
from plotly.matplotlylib import Exporter, PlotlyRenderer
from plotly.plotly import plot
import matplotlib.pyplot as plt

data = {
    'esN': [0, 1, 2, 3],
    'ewe_is0': [ -398.11901997, -398.11902774, -398.11897111, -398.11882215 ],
    'ewe_is1': [ -398.11793027, -398.11792966, -398.11786308, None ],
    'ewe_is2': [ -398.11397008, -398.11396421, None, None ]
}
df = pd.DataFrame.from_dict(data)
print type(df['ewe_is2'][3]), repr(df['ewe_is2'][3])

plotopts = { 'x': 'esN', 'marker': 'o' }
fig, ax = plt.subplots(1, 1)
df.plot(ax=ax, **plotopts)
for l in ax.get_lines(): print type(l.get_xydata()[-1][-1])

# following is debug version of plotly.plot_mpl()
# see https://github.com/plotly/python-api/blob/e5a3b3/plotly/tools.py#L370-L371
renderer = PlotlyRenderer()
Exporter(renderer).run(fig)
print type(renderer.plotly_fig['data'][1]['y'][-1])
print plot(renderer.plotly_fig, filename='test0_1', auto_open=False)