Pandas Secondary Axis

Question

I have the following data frame:

    Date        A           B
0   2017-05-31  17453139    5.865738
1   2017-06-30  17425164    5.272728
2   2017-07-31  17480789    4.843094

When I run this:

df.plot(x='Date', y='A')
df.B.plot(secondary_y=True)

I get the following error:

> appdata\local\programs\python\python36\lib\site-packages\pandas\plotting\_timeseries.py
> in format_dateaxis(subplot, freq, index)
>     335             TimeSeries_TimedeltaFormatter())
>     336     else:
> --> 337         raise TypeError('index type not supported')
>     338 
>     339     pylab.draw_if_interactive()
> 
> TypeError: index type not supported

And my graph looks like this underneath the error (blue and red should overlap):

Lanlana

Answer 1

Use secondary_y in pandas.DataFrame.plot:

ax = df.plot('Date', 'A')
df.plot('Date', 'B', secondary_y=True, ax=ax)

More succinctly:

ax = df.plot(x='Date', y=['A', 'B'], secondary_y='B', figsize=(8, 5))

# move the legend for each y axis
ax.legend(bbox_to_anchor=(1.05, 0.53), loc='center left', frameon=False)
ax.right_ax.legend(bbox_to_anchor=(1.05, 0.47), loc='center left', frameon=False)

Optionally, use ax.twinx()

ax = df.plot('Date','A')
ax1 = ax.twinx()
df.plot('Date','B',ax=ax1, color='r')

Output:

Stump

Answer 2

Cheat Sheet: Pandas plot ..all you need(Version v02, 30.12.22)

Variante 1: simply

df[["A", "B"]].plot(secondary_y=["B"])

Variante 2: with more settings

df[["A", "B"]].loc["2017"].plot(secondary_y=["B"], figsize=(14,7))
plt.title(r"my $\mathbf{The \hspace{0.1} Title}$") #..only bold single words with LaTeX ..\hspace{0.1} required for spacing between words
                                  #r"raw string" ..to ensure that the backslash character in TeX notation is not interpreted as an escape character
plt.show()

Variante 3: with additional information

year_start = "2010" #with pure zoom effect
year_end   = "2017" #
#a)left y-axis
df["A"].loc[year_start:year_end].plot(marker=".", linestyle="--", figsize=(14,7))
plt.ylabel("A-values")
plt.legend() #Legend ..from left y-axis, top right

#b)right y-axis
df["B"].loc[year_start:year_end].plot(secondary_y=["B"], color="blue", alpha=0.35)
plt.ylabel("B-values", fontweight="bold", fontstyle="italic") #draw whole text: bold and italic
plt.axhline(y=0, linestyle="--", label="zero line")
plt.axvline(x=df.index[splitInt], color="blue", linestyle="--", label="train-test-split")
plt.legend(loc ="center right") #Legend ..from right y-axis, "upper, center, lower + left, center, right"

plt.title(f"my Chart-Title: from {year_start} to {year_end}")
plt.tight_layout() #makes better use of chart space
ax = plt.gca() #you need an object ax and only there is the function set_facecolor() 
ax.set_facecolor("yellow") #set chart background color
plt.grid()
plt.show()

Solleret