"min() arg is an empty sequence" when trying to convert my pandas plot into plotly

Question

I receive an ValueError: min() arg is an empty sequence when trying to convert a pandas dataframe plot to plotly.

Here is a sample of the dataframe (with NaNs):

ttab.loc[:,:"Irish"]

Group   English (British)   Americans (White)   Canadians   Scots   Irish
Year                    
1926    1   2   3   4   5
1946    3   1   2   5   4
1956    3   1   2   7   5
1966    2   1   3   9   5
1977    2   1   3   9   7
1993    2   NaN NaN 6   1
2001    4   1   3   NaN 5
2001*   4   1   3   NaN 5
2012    4   1   3   NaN 6

I make a pretty plot:

# Time It
import time
begin = time.clock()

# Inline plotting
%matplotlib inline

# Import libraries
import matplotlib
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
from pandas.tools.plotting import scatter_matrix

# (*) To communicate with Plotly's server, sign in with credentials file
import plotly.plotly as py
# (*) Useful Python/Plotly tools
import plotly.tools as tls
# (*) Graph objects to piece together plots
from plotly.graph_objs import *

# Read, clean, slice data here...
# Rename columns, examine, rename groups here...
# Merge, set index, transpose dataframe here....
# Change datatype to float here...
# Define color palette here...
# Now Plot...

# Package all mpl plotting commands inside one function
def plot_mpl_fig():

    # plot parameters
    figsize = (12,12)
    axfontsize = 16
    legfontsize = 14
    labfontsize = 18
    yax = (30,0)
    lw = 2

    # Interpolate NaNs and plot as dashed lines
    ax = ttab.loc[:,:'Asian Indians'].interpolate().plot(color=colorlist, ylim=yax, lw=lw, fontsize=axfontsize, \
                                                     figsize=figsize, style='--')

    # Overplot the measured values with solid lines
    ttab.loc[:,:'Asian Indians'].plot(ax=ax, color=colorlist, ylim=yax, lw=lw, fontsize=axfontsize, figsize=figsize)

    # Legend handling
    lines, labels = ax.get_legend_handles_labels()
    ax.legend(lines[28:], labels[28:], loc='center left', bbox_to_anchor=(-0.3, 0.5), fontsize=legfontsize)

    # Labeling
    plt.xlabel('Year',fontsize=labfontsize)
    plt.ylabel('Rank',fontsize=labfontsize)

# Plot it!
plot_mpl_fig()

# N.B. get matplotlib figure object and assign a variable to it
mpl_fig1 = plt.figure()

And then I try to convert by following the example at: https://plot.ly/python/matplotlib-to-plotly-tutorial/

Quickstart:

>>> import plotly.plotly as py
>>> import matplotlib.pyplot as plt
>>> # auto sign-in with credentials or use py.sign_in()
>>> mpl_fig_obj= plt.figure()
>>> # ... make a matplotlib figure ...
>>> py.plot_mpl(mpl_fig1)                                  # <== ValueError

Or Tutorial at Section 6.1:

# Convert a matplotlib figure object to a Plotly figure object
help(tls.mpl_to_plotly)
py_fig1 = tls.mpl_to_plotly(mpl_fig1, verbose=True)         # <== ValueError
print(py_fig1.to_string())
# Plot a matplotlib figure in Plotly
help(py.iplot_mpl) 
py.iplot_mpl(mpl_fig1, filename='s6_damped_oscillation')    # <== ValueError

Anytime I call plot_mpl, mpl_to_plotly, or iplot_mpl I get the ValueError: min() arg is an empty sequence. I think it might have to do with the NaNs, but I don't know how to work around that. Any suggestions?

Answer 1

It looks like mpl_to_plotly requires using axes.plot(), not plt.plot(). The following code works in a Jupyter Notebook.

import plotly.offline as py
from plotly.offline import init_notebook_mode, iplot
import plotly.tools as tls
import matplotlib.pylab as plt
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas

init_notebook_mode(connected=True)
fig = plt.Figure()
ax = fig.gca()
x = [-2,0,4,6,7]
y = [q**2-q+3 for q in x]
ax.plot(x,y)
canvas = FigureCanvas(fig)
plotly_fig = tls.mpl_to_plotly(fig)
py.iplot(plotly_fig)