Plot 2-D Histogram in Python using Matplotlib
2D Histogram is used to analyze the relationship among two data variables which has wide range of values. A 2D histogram is very similar like 1D histogram. The class intervals of the data set are plotted on both x and y axis. Unlike 1D histogram, it drawn by including the total number of combinations of the values which occur in intervals of x and y, and marking the densities. It is useful when there is a large amount of data in a discrete distribution, and simplifies it by visualizing the points where the frequencies if variables are dense.
Creating a 2D Histogram
Matplotlib library provides an inbuilt function matplotlib.pyplot.hist2d() which is used to create 2D histogram.Below is the syntax of the function:
matplotlib.pyplot.hist2d(x, y, bins=(nx, ny), range=None, density=False, weights=None, cmin=None, cmax=None, cmap=value)
Here (x, y) specify the coordinates of the data variables, the length of the X data and Y variables should be same.The number of bins can be specified by the attribute bins=(nx, ny) where nx and ny is the number of bins to be used in the horizontal and vertical directions respectively.cmap=value is used to set the color scale.The range=None is an optional parameter used to set rectangular area in which data values are counted for plot.density=value is optional parameter accepting boolean values used to normalize histogram.
The code below code creates a simple 2D histogram using matplotlib.pyplot.hist2d() function having some random values of x and y:
# Import librariesimport numpy as npimport matplotlib.pyplot as pltimport random # Creating datasetn = 100x = np.random.standard_normal(n)y = 3.0 * x fig = plt.subplots(figsize =(10, 7))# Creating plotplot.hist2d(x, y)plot.title("Simple 2D Histogram") # show plotplot.show() |
Output:
Customizing 2D Histogram
The matplotlib.pyplot.hist2d() function has a wide range of methods which we can use to customize and create the plot for better view and understanding.
# Import librariesimport numpy as npimport matplotlib.pyplot as pltimport random # Creating datasetx = np.random.normal(size = 500000)y = x * 3 + 4 * np.random.normal(size = 500000) fig = plt.subplots(figsize =(10, 7))# Creating plotplot.hist2d(x, y)plot.title("Simple 2D Histogram") # show plotplot.show() |
Output:
Some of the customization of the above graph are listed below:
Changing the bin scale:-
# Import librariesimport numpy as npimport matplotlib.pyplot as pltimport random # Creating datasetx = np.random.normal(size = 500000)y = x * 3 + 4 * np.random.normal(size = 500000) # Creating binsx_min = np.min(x)x_max = np.max(x) y_min = np.min(y)y_max = np.max(y) x_bins = np.linspace(x_min, x_max, 50)y_bins = np.linspace(y_min, y_max, 20) fig, ax = plt.subplots(figsize =(10, 7))# Creating plotplt.hist2d(x, y, bins =[x_bins, y_bins])plt.title("Changing the bin scale") ax.set_xlabel('X-axis') ax.set_ylabel('X-axis') # show plotplt.tight_layout() plot.show() |
Output:
Changing the color scale and adding color bar:-
# Import librariesimport numpy as npimport matplotlib.pyplot as pltimport random # Creating datasetx = np.random.normal(size = 500000)y = x * 3 + 4 * np.random.normal(size = 500000) # Creating binsx_min = np.min(x)x_max = np.max(x) y_min = np.min(y)y_max = np.max(y) x_bins = np.linspace(x_min, x_max, 50)y_bins = np.linspace(y_min, y_max, 20) fig, ax = plt.subplots(figsize =(10, 7))# Creating plotplt.hist2d(x, y, bins =[x_bins, y_bins], cmap = plt.cm.nipy_spectral)plt.title("Changing the color scale and adding color bar") # Adding color barplt.colorbar() ax.set_xlabel('X-axis') ax.set_ylabel('X-axis') # show plotplt.tight_layout() plot.show() |
Output:
Filtering data:-
# Import librariesimport numpy as npimport matplotlib.pyplot as pltimport random # Creating datasetx = np.random.normal(size = 500000)y = x * 3 + 4 * np.random.normal(size = 500000) # Creating binsx_min = np.min(x)x_max = np.max(x) y_min = np.min(y)y_max = np.max(y) x_bins = np.linspace(x_min, x_max, 50)y_bins = np.linspace(y_min, y_max, 20) # Creating data filterdata = np.c_[x, y] for i in range(10000): x_idx = random.randint(0, 500000) data[x_idx, 0] = -9999 data = data[data[:, 0]!=-9999] fig, ax = plt.subplots(figsize =(10, 7))# Creating plotplt.hist2d(data[:, 0], data[:, 1], bins =[x_bins, y_bins])plt.title("Filtering data") ax.set_xlabel('X-axis') ax.set_ylabel('X-axis') # show plotplt.tight_layout() plot.show() |
Output:
Using matplotlib hexbin function:-
# Import librariesimport numpy as npimport matplotlib.pyplot as pltimport random # Creating datasetx = np.random.normal(size = 500000)y = x * 3 + 4 * np.random.normal(size = 500000) fig, ax = plt.subplots(figsize =(10, 7))# Creating plotplt.title("Using matplotlib hexbin function")plt.hexbin(x, y, bins = 50) ax.set_xlabel('X-axis') ax.set_ylabel('Y-axis') # show plotplt.tight_layout() plot.show() |
Output:


