I am a little confused about how this code works:
fig, axes = plt.subplots(nrows=2, ncols=2)
plt.show()
How does the fig, axes work in this case? What does it do?
Also why wouldn't this work to do the same thing:
fig = plt.figure()
axes = fig.subplots(nrows=2, ncols=2)
There are several ways to do it. The subplots method creates the figure along with the subplots that are then stored in the ax array. For example:
import matplotlib.pyplot as plt
x = range(10)
y = range(10)
fig, ax = plt.subplots(nrows=2, ncols=2)
for row in ax:
for col in row:
col.plot(x, y)
plt.show()
However, something like this will also work, it's not so "clean" though since you are creating a figure with subplots and then add on top of them:
fig = plt.figure()
plt.subplot(2, 2, 1)
plt.plot(x, y)
plt.subplot(2, 2, 2)
plt.plot(x, y)
plt.subplot(2, 2, 3)
plt.plot(x, y)
plt.subplot(2, 2, 4)
plt.plot(x, y)
plt.show()
import matplotlib.pyplot as plt
fig, ax = plt.subplots(2, 2)
ax[0, 0].plot(range(10), 'r') #row=0, col=0
ax[1, 0].plot(range(10), 'b') #row=1, col=0
ax[0, 1].plot(range(10), 'g') #row=0, col=1
ax[1, 1].plot(range(10), 'k') #row=1, col=1
plt.show()
You can also unpack the axes in the subplots call
And set whether you want to share the x and y axes between the subplots
Like this:
import matplotlib.pyplot as plt
# fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(nrows=2, ncols=2, sharex=True, sharey=True)
fig, axes = plt.subplots(nrows=2, ncols=2, sharex=True, sharey=True)
ax1, ax2, ax3, ax4 = axes.flatten()
ax1.plot(range(10), 'r')
ax2.plot(range(10), 'b')
ax3.plot(range(10), 'g')
ax4.plot(range(10), 'k')
plt.show()
You might be interested in the fact that as of matplotlib version 2.1 the second code from the question works fine as well.
From the change log:
Figure class now has subplots method
The Figure class now has a subplots() method which behaves the same as pyplot.subplots() but on an existing figure.
Example:
import matplotlib.pyplot as plt
fig = plt.figure()
axes = fig.subplots(nrows=2, ncols=2)
plt.show()
Read the documentation: matplotlib.pyplot.subplots
pyplot.subplots() returns a tuple fig, ax which is unpacked in two variables using the notation
fig, axes = plt.subplots(nrows=2, ncols=2)
The code:
fig = plt.figure()
axes = fig.subplots(nrows=2, ncols=2)
does not work because subplots() is a function in pyplot not a member of the object Figure.
Iterating through all subplots sequentially:
fig, axes = plt.subplots(nrows, ncols)
for ax in axes.flatten():
ax.plot(x,y)
Accessing a specific index:
for row in range(nrows):
for col in range(ncols):
axes[row,col].plot(x[row], y[col])
Subplots with pandas
This answer is for subplots with pandas, which uses matplotlib as the default plotting backend.
Here are four options to create subplots starting with a pandas.DataFrame
Implementation 1. and 2. are for the data in a wide format, creating subplots for each column.
Implementation 3. and 4. are for data in a long format, creating subplots for each unique value in a column.
Tested in python 3.8.11, pandas 1.3.2, matplotlib 3.4.3, seaborn 0.11.2
Imports and Data
import seaborn as sns # data only
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
# wide dataframe
df = sns.load_dataset('planets').iloc[:, 2:5]
orbital_period mass distance
0 269.300 7.10 77.40
1 874.774 2.21 56.95
2 763.000 2.60 19.84
3 326.030 19.40 110.62
4 516.220 10.50 119.47
# long dataframe
dfm = sns.load_dataset('planets').iloc[:, 2:5].melt()
variable value
0 orbital_period 269.300
1 orbital_period 874.774
2 orbital_period 763.000
3 orbital_period 326.030
4 orbital_period 516.220
1. subplots=True and layout, for each column
Use the parameters subplots=True and layout=(rows, cols) in pandas.DataFrame.plot
This example uses kind='density', but there are different options for kind, and this applies to them all. Without specifying kind, a line plot is the default.
ax is array of AxesSubplot returned by pandas.DataFrame.plot
See How to get a Figure object, if needed.
How to save pandas subplots
axes = df.plot(kind='density', subplots=True, layout=(2, 2), sharex=False, figsize=(10, 6))
# extract the figure object; only used for tight_layout in this example
fig = axes[0][0].get_figure()
# set the individual titles
for ax, title in zip(axes.ravel(), df.columns):
ax.set_title(title)
fig.tight_layout()
plt.show()
2. plt.subplots, for each column
Create an array of Axes with matplotlib.pyplot.subplots and then pass axes[i, j] or axes[n] to the ax parameter.
This option uses pandas.DataFrame.plot, but can use other axes level plot calls as a substitute (e.g. sns.kdeplot, plt.plot, etc.)
It's easiest to collapse the subplot array of Axes into one dimension with .ravel or .flatten. See .ravel vs .flatten.
Any variables applying to each axes, that need to be iterate through, are combined with .zip (e.g. cols, axes, colors, palette, etc.). Each object must be the same length.
fig, axes = plt.subplots(nrows=2, ncols=2, figsize=(10, 6)) # define the figure and subplots
axes = axes.ravel() # array to 1D
cols = df.columns # create a list of dataframe columns to use
colors = ['tab:blue', 'tab:orange', 'tab:green'] # list of colors for each subplot, otherwise all subplots will be one color
for col, color, ax in zip(cols, colors, axes):
df[col].plot(kind='density', ax=ax, color=color, label=col, title=col)
ax.legend()
fig.delaxes(axes[3]) # delete the empty subplot
fig.tight_layout()
plt.show()
Result for 1. and 2.
3. plt.subplots, for each group in .groupby
This is similar to 2., except it zips color and axes to a .groupby object.
fig, axes = plt.subplots(nrows=2, ncols=2, figsize=(10, 6)) # define the figure and subplots
axes = axes.ravel() # array to 1D
dfg = dfm.groupby('variable') # get data for each unique value in the first column
colors = ['tab:blue', 'tab:orange', 'tab:green'] # list of colors for each subplot, otherwise all subplots will be one color
for (group, data), color, ax in zip(dfg, colors, axes):
data.plot(kind='density', ax=ax, color=color, title=group, legend=False)
fig.delaxes(axes[3]) # delete the empty subplot
fig.tight_layout()
plt.show()
4. seaborn figure-level plot
Use a seaborn figure-level plot, and use the col or row parameter. seaborn is a high-level API for matplotlib. See seaborn: API reference
p = sns.displot(data=dfm, kind='kde', col='variable', col_wrap=2, x='value', hue='variable',
facet_kws={'sharey': False, 'sharex': False}, height=3.5, aspect=1.75)
sns.move_legend(p, "upper left", bbox_to_anchor=(.55, .45))
Convert the axes array to 1D
Generating subplots with plt.subplots(nrows, ncols), where both nrows and ncols is greater than 1, returns a nested array of <AxesSubplot:> objects.
It’s not necessary to flatten axes in cases where either nrows=1 or ncols=1, because axes will already be 1 dimensional, which is a result of the default parameter squeeze=True
The easiest way to access the objects, is to convert the array to 1 dimension with .ravel(), .flatten(), or .flat.
.ravel vs. .flatten
flatten always returns a copy.
ravel returns a view of the original array whenever possible.
Once the array of axes is converted to 1-d, there are a number of ways to plot.
This answer is relevant to seaborn axes-level plots, which have the ax= parameter (e.g. sns.barplot(…, ax=ax[0]).
seaborn is a high-level API for matplotlib. See Figure-level vs. axes-level functions and seaborn is not plotting within defined subplots
import matplotlib.pyplot as plt
import numpy as np # sample data only
# example of data
rads = np.arange(0, 2*np.pi, 0.01)
y_data = np.array([np.sin(t*rads) for t in range(1, 5)])
x_data = [rads, rads, rads, rads]
# Generate figure and its subplots
fig, axes = plt.subplots(nrows=2, ncols=2)
# axes before
array([[<AxesSubplot:>, <AxesSubplot:>],
[<AxesSubplot:>, <AxesSubplot:>]], dtype=object)
# convert the array to 1 dimension
axes = axes.ravel()
# axes after
array([<AxesSubplot:>, <AxesSubplot:>, <AxesSubplot:>, <AxesSubplot:>],
dtype=object)
Iterate through the flattened array
If there are more subplots than data, this will result in IndexError: list index out of range
Try option 3. instead, or select a subset of the axes (e.g. axes[:-2])
for i, ax in enumerate(axes):
ax.plot(x_data[i], y_data[i])
Access each axes by index
axes[0].plot(x_data[0], y_data[0])
axes[1].plot(x_data[1], y_data[1])
axes[2].plot(x_data[2], y_data[2])
axes[3].plot(x_data[3], y_data[3])
Index the data and axes
for i in range(len(x_data)):
axes[i].plot(x_data[i], y_data[i])
zip the axes and data together and then iterate through the list of tuples.
for ax, x, y in zip(axes, x_data, y_data):
ax.plot(x, y)
Ouput
An option is to assign each axes to a variable, fig, (ax1, ax2, ax3) = plt.subplots(1, 3). However, as written, this only works in cases with either nrows=1 or ncols=1. This is based on the shape of the array returned by plt.subplots, and quickly becomes cumbersome.
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2) for a 2 x 2 array.
This option is most useful for two subplots (e.g.: fig, (ax1, ax2) = plt.subplots(1, 2) or fig, (ax1, ax2) = plt.subplots(2, 1)). For more subplots, it's more efficient to flatten and iterate through the array of axes.
You could use the following:
import numpy as np
import matplotlib.pyplot as plt
fig, _ = plt.subplots(nrows=2, ncols=2)
for i, ax in enumerate(fig.axes):
ax.plot(np.sin(np.linspace(0,2*np.pi,100) + np.pi/2*i))
Or alternatively, using the second variable that plt.subplot returns:
fig, ax_mat = plt.subplots(nrows=2, ncols=2)
for i, ax in enumerate(ax_mat.flatten()):
...
ax_mat is a matrix of the axes. It's shape is nrows x ncols.
here is a simple solution
fig, ax = plt.subplots(nrows=2, ncols=3, sharex=True, sharey=False)
for sp in fig.axes:
sp.plot(range(10))
Go with the following if you really want to use a loop:
def plot(data):
fig = plt.figure(figsize=(100, 100))
for idx, k in enumerate(data.keys(), 1):
x, y = data[k].keys(), data[k].values
plt.subplot(63, 10, idx)
plt.bar(x, y)
plt.show()
Another concise solution is:
// set up structure of plots
f, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(20,10))
// for plot 1
ax1.set_title('Title A')
ax1.plot(x, y)
// for plot 2
ax2.set_title('Title B')
ax2.plot(x, y)
// for plot 3
ax3.set_title('Title C')
ax3.plot(x,y)
I am plotting multiple lines on a single plot and I want them to run through the spectrum of a colormap, not just the same 6 or 7 colors. The code is akin to this:
for i in range(20):
for k in range(100):
y[k] = i*x[i]
plt.plot(x,y)
plt.show()
Both with colormap "jet" and another that I imported from seaborn, I get the same 7 colors repeated in the same order. I would like to be able to plot up to ~60 different lines, all with different colors.
The Matplotlib colormaps accept an argument (0..1, scalar or array) which you use to get colors from a colormap. For example:
col = pl.cm.jet([0.25,0.75])
Gives you an array with (two) RGBA colors:
array([[ 0. , 0.50392157, 1. , 1. ],
[ 1. , 0.58169935, 0. , 1. ]])
You can use that to create N different colors:
import numpy as np
import matplotlib.pylab as pl
x = np.linspace(0, 2*np.pi, 64)
y = np.cos(x)
pl.figure()
pl.plot(x,y)
n = 20
colors = pl.cm.jet(np.linspace(0,1,n))
for i in range(n):
pl.plot(x, i*y, color=colors[i])
Bart's solution is nice and simple but has two shortcomings.
plt.colorbar() won't work in a nice way because the line plots aren't mappable (compared to, e.g., an image)
It can be slow for large numbers of lines due to the for loop (though this is maybe not a problem for most applications?)
These issues can be addressed by using LineCollection. However, this isn't too user-friendly in my (humble) opinion. There is an open suggestion on GitHub for adding a multicolor line plot function, similar to the plt.scatter(...) function.
Here is a working example I was able to hack together
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.collections import LineCollection
def multiline(xs, ys, c, ax=None, **kwargs):
"""Plot lines with different colorings
Parameters
----------
xs : iterable container of x coordinates
ys : iterable container of y coordinates
c : iterable container of numbers mapped to colormap
ax (optional): Axes to plot on.
kwargs (optional): passed to LineCollection
Notes:
len(xs) == len(ys) == len(c) is the number of line segments
len(xs[i]) == len(ys[i]) is the number of points for each line (indexed by i)
Returns
-------
lc : LineCollection instance.
"""
# find axes
ax = plt.gca() if ax is None else ax
# create LineCollection
segments = [np.column_stack([x, y]) for x, y in zip(xs, ys)]
lc = LineCollection(segments, **kwargs)
# set coloring of line segments
# Note: I get an error if I pass c as a list here... not sure why.
lc.set_array(np.asarray(c))
# add lines to axes and rescale
# Note: adding a collection doesn't autoscalee xlim/ylim
ax.add_collection(lc)
ax.autoscale()
return lc
Here is a very simple example:
xs = [[0, 1],
[0, 1, 2]]
ys = [[0, 0],
[1, 2, 1]]
c = [0, 1]
lc = multiline(xs, ys, c, cmap='bwr', lw=2)
Produces:
And something a little more sophisticated:
n_lines = 30
x = np.arange(100)
yint = np.arange(0, n_lines*10, 10)
ys = np.array([x + b for b in yint])
xs = np.array([x for i in range(n_lines)]) # could also use np.tile
colors = np.arange(n_lines)
fig, ax = plt.subplots()
lc = multiline(xs, ys, yint, cmap='bwr', lw=2)
axcb = fig.colorbar(lc)
axcb.set_label('Y-intercept')
ax.set_title('Line Collection with mapped colors')
Produces:
Hope this helps!
An anternative to Bart's answer, in which you do not specify the color in each call to plt.plot is to define a new color cycle with set_prop_cycle. His example can be translated into the following code (I've also changed the import of matplotlib to the recommended style):
import numpy as np
import matplotlib.pyplot as plt
x = np.linspace(0, 2*np.pi, 64)
y = np.cos(x)
n = 20
ax = plt.axes()
ax.set_prop_cycle('color',[plt.cm.jet(i) for i in np.linspace(0, 1, n)])
for i in range(n):
plt.plot(x, i*y)
If you are using continuous color pallets like brg, hsv, jet or the default one then you can do like this:
color = plt.cm.hsv(r) # r is 0 to 1 inclusive
Now you can pass this color value to any API you want like this:
line = matplotlib.lines.Line2D(xdata, ydata, color=color)
This approach seems to me like the most concise, user-friendly and does not require a loop to be used. It does not rely on user-made functions either.
import numpy as np
import matplotlib.pyplot as plt
# make 5 lines
n_lines = 5
x = np.arange(0, 2).reshape(-1, 1)
A = np.linspace(0, 2, n_lines).reshape(1, -1)
Y = x # A
# create colormap
cm = plt.cm.bwr(np.linspace(0, 1, n_lines))
# plot
ax = plt.subplot(111)
ax.set_prop_cycle('color', list(cm))
ax.plot(x, Y)
plt.show()
Resulting figure here
Background:
I have a list_of_x_and_y_list that contains x and y values which looks like:
[[(44800, 14888), (132000, 12500), (40554, 12900)], [(None, 193788), (101653, 78880), (3866, 160000)]]
I have another data_name_list ["data_a","data_b"] so that
"data_a" = [(44800, 14888), (132000, 12500), (40554, 12900)]
"data_b" = [(None, 193788), (101653, 78880), (3866, 160000)]
The len of list_of_x_and_y_list / or len of data_name_list is > 20.
Question:
How can I create a scatter plot for each item (being the same colour) in the data_name_list?
What I have tried:
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax = plt.axes(facecolor='#FFFFFF')
prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']
print(list_of_x_and_y_list)
for x_and_y_list, data_name, color in zip(list_of_x_and_y_list, data_name_list, colors):
for x_and_y in x_and_y_list,:
print(x_and_y)
x, y = x_and_y
ax.scatter(x, y, label=data_name, color=color) # "label=data_name" creates
# a huge list as a legend!
# :(
plt.title('Matplot scatter plot')
plt.legend(loc=2)
file_name = "3kstc.png"
fig.savefig(file_name, dpi=fig.dpi)
print("Generated: {}".format(file_name))
The Problem:
The legend appears to be a very long list, which I don't know how to rectify:
Relevant Research:
Matplotlib scatterplot
Scatter Plot
Scatter plot in Python using matplotlib
The reason you get a long repeated list as a legend is because you are providing each point as a separate series, as matplotlib does not automatically group your data based on the labels.
A quick fix is to iterate over the list and zip together the x-values and the y-values of each series as two tuples, so that the x tuple contains all the x-values and the y tuple the y-values.
Then you can feed these tuples to the plt.plot method together with the labels.
I felt that the names list_of_x_and_y_list were uneccessary long and complicated, so in my code I've used shorter names.
import matplotlib.pyplot as plt
data_series = [[(44800, 14888), (132000, 12500), (40554, 12900)],
[(None, 193788), (101653, 78880), (3866, 160000)]]
data_names = ["data_a","data_b"]
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax = plt.axes(facecolor='#FFFFFF')
prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']
for data, data_name, color in zip(data_series, data_names, colors):
x,y = zip(*data)
ax.scatter(x, y, label=data_name, color=color)
plt.title('Matplot scatter plot')
plt.legend(loc=1)
To only get one entry per data_name, you should add data_name only once as a label. The rest of the calls should go with label=None.
The simplest you can achieve this using the current code, is to set data_name to None at the end of the loop:
from matplotlib import pyplot as plt
from random import randint
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax.set_facecolor('#FFFFFF')
# create some random data, suppose the sublists have different lengths
list_of_x_and_y_list = [[(randint(1000, 4000), randint(2000, 5000)) for col in range(randint(2, 10))]
for row in range(10)]
data_name_list = list('abcdefghij')
colors = plt.rcParams['axes.prop_cycle'].by_key()['color']
for x_and_y_list, data_name, color in zip(list_of_x_and_y_list, data_name_list, colors):
for x_and_y in x_and_y_list :
x, y = x_and_y
ax.scatter(x, y, label=data_name, color=color)
data_name = None
plt.legend(loc=2)
plt.show()
Some things can be simplified, making the code 'more pythonic', for example:
for x_and_y in x_and_y_list :
x, y = x_and_y
can be written as:
for x, y in x_and_y_list:
Another issue, is that with a lot of data calling scatter for every point could be rather slow. All the x and y belonging to the same list can be plotted together. For example using list comprehension:
for x_and_y_list, data_name, color in zip(list_of_x_and_y_list, data_name_list, colors):
xs = [x for x, y in x_and_y_list]
ys = [y for x, y in x_and_y_list]
ax.scatter(xs, ys, label=data_name, color=color)
scatter could even get a list of colors per point, but plotting all the points in one go, wouldn't allow for labels per data_name.
Very often, numpy is used to store numerical data. This has some advantages, such as vectorization for quick calculations. With numpy the code would look like:
import numpy as np
for x_and_y_list, data_name, color in zip(list_of_x_and_y_list, data_name_list, colors):
xys = np.array(x_and_y_list)
ax.scatter(xys[:,0], xys[:,1], label=data_name, color=color)
I want to plot only positive values when plotting a graph (like the RELU function in ML)
This may well be a dumb question. I hope not.
In the code below I iterate and change the underlying list data. I really want to only change the values when it's plot time and not change the source list data. Is that possible?
#create two lists in range -10 to 10
x = list(range(-10, 11))
y = list(range(-10, 11))
#this function changes the underlying data to remove negative values
#I really want to do this at plot time
#I don't want to change the source list. Can it be done?
for idx, val in enumerate(y):
y[idx] = max(0, val)
#a bunch of formatting to make the plot look nice
plt.figure(figsize=(6, 6))
plt.axhline(y=0, color='silver')
plt.axvline(x=0, color='silver')
plt.grid(True)
plt.plot(x, y, 'rx')
plt.show()
I'd suggest using numpy and filter the data when plotting:
import numpy as np
import matplotlib.pyplot as plt
#create two lists in range -10 to 10
x = list(range(-10, 11))
y = list(range(-10, 11))
x = np.array(x)
y = np.array(y)
#a bunch of formatting to make the plot look nice
plt.figure(figsize=(6, 6))
plt.axhline(y=0, color='silver')
plt.axvline(x=0, color='silver')
plt.grid(True)
# plot only those values where y is positive
plt.plot(x[y>0], y[y>0], 'rx')
plt.show()
This will not plot points with y < 0 at all. If instead, you want to replace any negative value by zero, you can do so as follows
plt.plot(x, np.maximum(0,y), 'rx')
It may look a bit complicated but filter the data on the fly:
plt.plot(list(zip(*[(x1,y1) for (x1,y1) in zip(x,y) if x1>0])), 'rx')
Explanation: it is safer to handle the data as pairs so that (x,y) stay in sync, and then you have to convert pairs back to separate xlist and ylist.