[PYTHON] Grammar summary that is often forgotten with matplotlib

This is a memo that is often forgotten when dealing with the Python drawing library "matplotlib".

• Change log

2016-11-11 「ax.Added "annotate characters and arrows" and fine-tuned headings
2016-11-11 Added "Locator scale position"
2016-12-16 Updated "Locator Scale Position"
2017-02-11 Added "out of frame" method for legend

--Environment

- Anaconda 4.0.0 (Python 2.7.13)
- CentOS 6.5

--Scheduled to be added in the future (2016-11-11)

- cmap
- plt.xlim() = ax.set_xlim()

** Tend to forget plot **

** plt.fill_between ** fill

Ranged fill plot

plt.fill_between(time, value1, value2)

plt.fill_betweenx(depth, profile1, profile2) #When writing a vertical distribution

http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.fill_between

** plt.errorbar ** error bar

Plot with error bars

plt.errorbar(time, value, yerr=value_std)

plt.errorbar(profile, depth, xerr=profile_std) #For vertical distribution

http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.errorbar

** df.plot ** Plot for pandas

Plot of pandas.DataFrame

df.plot(x='time', y='value1', ls='-', marker='o', color='k')

df.plot(x='time', y=['value1','value2']) #Draw multiple lines on one graph.

df.plot(x='time', y=['value1','value2'], subplots=True, layout=(1,2)) # ax[0], ax[1]Depicted in each

--x is ʻindex by default, so it can be omitted, but I don't know how to specify it as y. (Y ='index'` will result in an error) --y Omission plots all columns

http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.plot.html

** plt.semilogy ** Logarithmic axis plot

plt.semilogy(x, y) #y-axis is logarithmic

plt.semilogx(x, y) #x-axis is logarithmic

plt.loglog(x, y) #Log-log

ax.plot(x, y)
ax.set_xscale("log") #Can be set later

ax.set_yscale("log", nonposy='clip') #What to do if it becomes negative.'mask'There is also

http://matplotlib.org/examples/pylab_examples/log_demo.html http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.semilogy

** ax.annotate ** letters and arrows

ax.annotate('arrowstyle', 
            xy=(0, 1), xycoords='data',      #The tip of the arrow(xy)Axle value(data)Specified by
            xytext=(-50, 30), 
            textcoords='offset points',      #Character position(xytext)The relative distance from the tip of the arrow(offset points)Specified by
            arrowprops=dict(arrowstyle="->")
            )

http://matplotlib.org/examples/pylab_examples/annotation_demo2.html

** Tend to forget axis setting **

** ax.set_xticklabels ([]) ** Erase labels

Erase the axis label (number or date part)

fig, ax = plt.subplots(1,2)

ax[0].set_xticklabels([]) #Erase the x-axis in the figure above

** DateFormatter ** Change date format

Change the time axis format

from matplotlib.dates import DateFormatter

ax.xaxis.set_major_formatter(DateFormatter('%m/%d\n%H:%M'))

http://matplotlib.org/api/dates_api.html#matplotlib.dates.DateFormatter

** Locator ** Scale position

** Value axis **

from matplotlib import ticker

ax.xaxis.set_major_locator(ticker.MultipleLocator(20))           #Every 20
ax.xaxis.set_major_locator(ticker.MaxNLocator(5))                #Up to 5

** Time axis **

from matplotlib import dates as mdates

ax.xaxis.set_major_locator(mdates.AutoDateLocator(maxticks=8))   #Up to 8
ax.xaxis.set_major_locator(mdates.HourLocator(12))               #12 o'clock
ax.xaxis.set_major_locator(mdates.HourLocator(interval=24))      #Every 24 hours
ax.xaxis.set_major_locator(mdates.DayLocator(np.arange(1,31,7))) #Weekly

** Forgotten options **

** plt.savefig ** Save

http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.savefig

** Background transparency **

plt.savefig(filename, transparent=True) # default: False

** Small margins **

plt.savefig(filename, bbox_inches='tight')

** plt.legend ** Legend

http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.legend

** Change the arrangement **

plt.legend(ncol=2) #The number of each legend side by side (default): 1）

** Number of markers **

plt.legend(numpoints=1) #Unify the number of markers (default):For some reason 2)

Match legend points with matplotlib @ halm

** Put out of the frame **

plt.legend(bbox_to_anchor=(1.01,1), loc=2, borderaxespad=0)

http://matplotlib.org/users/legend_guide.html#legend-location

** Other **

** plt.rcParams [] ** Settings

If you set the font etc. before drawing, it will be quick.

plt.rcParams["font.size"] = 12 #Change font size (default): 10?）

http://matplotlib.org/users/customizing.html

** matplotlibrc ** default settings

Or it's easier to change the default settings, so edit matplotlibrc

matplotlibrc

legend.numpoints = 1

** plt.style.use ** style

Overall design changes

plt.style.use('ggplot')

Reference page

-Personal note: Adjusting the graph using matplotlib @ hysterect -[Python] Customize Colormap when drawing graphs with matplotlib @ kenmatsu4