`mpl deprecation warning` for converting simple `plt.scatterplot` to `bokeh`

import bokeh
print(bokeh.__version__)
#0.12.2

# Plot from Matplotlib

# Tools (Not sure how to add this when converting from `matplotlib` or `Seaborn`)
hover = HoverTool(
        tooltips=[
            ("index", "@index"),
            ("(x,y)", "(@t, $y)"),
            ("desc", "@desc"),
        ]
    )

# Data
DF_data = pd.DataFrame([a, np.sin(np.pi*a), np.cos(np.pi*a)], index=["t", "sine", "cosine"], columns=["t_%d"%_ for _ in range(100)]).T
DF_data["desc"] = ["info about this" for _ in DF_data.index]

# Matplotlib
fig, ax = plt.subplots(figsize=(15,5))
ax.scatter(data=DF_data, x="t", y="sine", s=np.linspace(0,100,100), alpha=0.5, color="green")
ax.scatter(data=DF_data, x="t", y="cosine", s=np.linspace(0,100,100)[::-1], alpha=0.5, color="teal")

# Convert to Bokeh
bokeh.io.show(bokeh.mpl.to_bokeh())

The error:

/Users/Mu/anaconda/lib/python3.5/site-packages/bokeh/core/compat/bokeh_renderer.py:257: UserWarning:

Path marker shapes currently not handled, defaulting to Circle

/Users/Mu/anaconda/lib/python3.5/site-packages/matplotlib/artist.py:221: MatplotlibDeprecationWarning:

This has been deprecated in mpl 1.5, please use the axes property. A removal date has not been set.

/Users/Mu/anaconda/lib/python3.5/site-packages/bokeh/core/compat/bokeh_renderer.py:290: UserWarning:

Path marker sizes support is limited and may not display as expected

The output plot:

[bryevdv]

Scatter works best with “tidy” data frames, so if you had a “type” column that had values “sine” or “cosine” for each t and y, you could do: Scatter(df, x="t", y="y", color="type")

Edit: pd.melt can do this transformation: http://pandas.pydata.org/pandas-docs/stable/generated/pandas.melt.html

[bryevdv]

@jolespin yes if you need to control the alpha then currently you should definitely stick with bokeh.plotting. It would not be too difficult (I don’t think) to add support for an alpha setting on some charts, but it can’t be a high priority at the moment given all that there is to do. If you are interested in looking into it, I an point you in the right direction.

I should also mention that there are some improvements coming in 0.12.3 that make mapping colors and alphas directly to data columns much easier with bokeh.plotting. I don’t have time to elaborate this morning but you can check recent PRs, or await the release announcement.

I am going to close this issue, since the discussion has turned to support type questions. Those are more properly directed to the mailing list: https://groups.google.com/a/continuum.io/forum/?pli=1#!forum/bokeh