Data Visualization in Python using Bokeh Library

Bokeh is a Python library which is used for data visualization through high-performance interactive charts and plots. It creates its plots using HTML and JavaScript languages. The output of the bokeh library can be generated on several platforms such as browser, HTML, server, and notebook. It is also possible to create the bokeh plots in Django and flask applications.

The bokeh library provides two visualization interfaces to the users:

models: it is a low-level interface which provides high flexibility to the application developers.
plotting: it is a high-level interface which is used for creating visual glyphs.

In this tutorial, we will learn how to create different types of data visualization graphs and charts using the bokeh library in Python.

Installation:

To install the bokeh library, we can use the following command:

Code 1: To Create Scatter Circle Markers

We can create Scatter Circle Markers on the Plot by using the bokeh library. For this, we will use the circle() function.

Example:

  # First, we will import the required modules  from bokeh.plotting import figure as fig  from bokeh.plotting import output_notebook as ON  from bokeh.plotting import show      # output to the notebook  ON()      # creating figure  plot1 = fig(plot_width = 500, plot_height = 500, title = “Scatter Markers”)      # adding a circle renderer with size, color and alpha  plot1.circle([1, 2, 3, 4, 5, 6], [2, 1, 6, 8, 0],            size = 12, color = “green”, alpha = 1)      # show the results  show(plot1)  

Output:

Code 2: To Create a Single Line

We can create a single line by using the bokeh library in Python. For this, we will use the line() function.

Example:

  # First, we will import the required modules  from bokeh.plotting import figure as fig  from bokeh.plotting import output_notebook as ON  from bokeh.plotting import show      # output to the notebook  ON()      # creating figure  plot1 = fig(plot_width = 500, plot_height = 500, title = “Scatter Markers”)       # adding a line renderer  plot1.line([1, 2, 3, 4, 5, 6], [2, 1, 6, 8, 0],           line_width = 4, color = “red”)      # show the results  show(plot1)  

Output:

Code 3: To Create a Bar chart

Bar charts are used for representing the categorical data with rectangular bars. The length of the bar is proportional to the values they are representing.

Example: Vertical Bar Chart

  # First, we will import the required modules  from bokeh.plotting import figure as fig  from bokeh.plotting import output_file as OF  from bokeh.plotting import show       # file to save the model  output_file(“jtp.html”)           # Now, instantiate the figure object  graph1 = fig(title = “Bokeh Bar Graph”)       # x-coordinates to be plotted  x = [1, 2, 3, 4, 5, 6, 7, 8]       # y-coordinates of the top edges  top1 = [5, 1, 4, 3, 2, 7, 6, 8]       # width / thickness of the bars   width1 = 0.7       # plot the graph  graph1.vbar(x,             top = top1,             width = width1,             color = “green”)       # display the model  show(graph1)  

Output:

Example 2: Horizontal Bar Chart

  # First, we will import the required modules  from bokeh.plotting import figure as fig  from bokeh.plotting import output_file as OF  from bokeh.plotting import show       # file to save the model  output_file(“jtp.html”)           # Now, instantiate the figure object  graph1 = fig(title = “Bokeh Bar Graph”)       # x-coordinates to be plotted  x = [1, 2, 3, 4, 5, 6, 7, 8]       # y-coordinates of the top edges  right1 = [5, 1, 4, 3, 2, 7, 6, 8]       # width / thickness of the bars   height1 = 0.7       # plot the graph  graph1.hbar(x,             right = right1,             height = height1,             color = “green”)       # display the model  show(graph1)  

Output:

Code 4: To Create Box Plot

Box plot is used for representing the statistical data on the plots, and it is helpful for summarizing the statistical properties of several groups present in the dataset.

Example:

  import numpy as nmp  import pandas as pnd    from bokeh.plotting import figure as fig  from bokeh.plotting import show    # generate some time series for six different categories  cats_1 = list(“xyzprq”)  y_y = nmp.random.randn(2000)  g_1 = nmp.random.choice(cats_1, 2000)  for K, J in enumerate(cats_1):      y_y[g_1 == l] += K // 2  data_frame = pnd.DataFrame(dict(score = y_y, group = g_1))    # now, we will find the quartiles and IQR for each category  groups1 = data_frame.groupby(‘group’)  q_1 = groups1.quantile(q = 0.25)  q_2 = groups1.quantile(q = 0.5)  q_3 = groups1.quantile(q = 0.75)  iqr = q3 – q1  upper = q3 + 1.5 * iqr  lower = q1 – 1.5 * iqr    # now we will find the outliers for each category  def outliers(group2):      cat_2 = group2.name      return group2[(group2.score > upper.loc[cat_2][‘score’]) | (group2.score < lower.loc[cat_2][‘score’])][‘score’]  out = groups.apply(outliers).dropna()    # here, we will prepare outlier data to plot, we would be needing coordinates for every outlier.  if not out.empty:      out_x = list(out.index.get_level_values(0))      out_y = list(out.values)    plot1 = fig(tools = “”, background_fill_color = “#efefef”, x_range = cats_1, toolbar_location = None)    # if no outliers, we will shrink the lengths of stems to be no longer than the minimums or maximums  q_min = groups1.quantile(q = 0.00)  q_max = groups1.quantile(q = 1.00)  upper.score = [min([x,y]) for (x,y) in zip(list(q_max.loc[:, ‘score’]), upper.score)]  lower.score = [max([x,y]) for (x,y) in zip(list(q_min.loc[:, ‘score’]), lower.score)]    # stems  plot1.segment(cats_1, upper.score, cats_1, q_3.score, line_color = “black”)  plot1.segment(cats_1, lower.score, cats_1, q_1.score, line_color = “black”)    # boxes  plot1.vbar(cats_1, 0.7, q_2.score, q_3.score, fill_color = “#E08E79”, line_color = “black”)  plot1.vbar(cats_1, 0.7, q_1.score, q_2.score, fill_color = “#3B8686”, line_color = “black”)    # whiskers (almost-0 height rects simpler than segments)  plot1.rect(cats_1, lower.score, 0.2, 0.01, line_color = “black”)  plot1.rect(cats_1, upper.score, 0.2, 0.01, line_color = “black”)    # outliers  if not out.empty:      plot1.circle(outx, outy, size = 6, color = “#F38630”, fill_alpha = 0.6)    plot1.xgrid.grid_line_color = None  plot1.ygrid.grid_line_color = “white”  plot1.grid.grid_line_width = 2  plot1.xaxis.major_label_text_font_size = “16px”    show(plot1)  

Output:

Code 5: To Create Scatter Plot

A Scatter plot is used for plotting values of two variables in a dataset. It is helpful in finding the correlation between the two variables which are selected.

Example:

  # First, we will import the required modules  from bokeh.plotting import figure as fig  from bokeh.plotting import output_file as OF  from bokeh.plotting import show       # file for saving the model  OF(“jtp.html”)            # instantiate the figure object  graph1 = fig(title = “Bokeh Scatter Graph”)          # the points to be plotted on Scatter Plot  x1 = [1.4, 5.1, 5.9, 2.3, 5.6, 8.6, 4.5, 2.1, 3.1, 4.3, 5.5, 4.4, 6.9, 2.1, 4, 5.2, 6.3, 7.2, 7.9, 2]  y1 = [3.4, 2.1, 5.7, 8.5, 4.3, 4.2, 5.7, 6.5, 8.9, 9.1, 1.5, 2.1, 6.8, 1, 6, 5.2, 4.5, 7.4, 7.5, 6.3]         # plott the graph  graph1.scatter(x1, y1)          # display the model  show(graph1)  

Output:

Conclusion

In this tutorial, we have discussed various types of data visualization, namely Bar chart, Box plot, Scatter Markers, Scatter Plots, and Single lines, which can be created using the bokeh library in Python.

Next TopicHow to Plot glyphs over a Google Map by using Bokeh Library in Python

Data Visualization in Python using Bokeh Library

Data Visualization in Python using Bokeh Library

Installation:

Code 1: To Create Scatter Circle Markers

Code 2: To Create a Single Line

Code 3: To Create a Bar chart

Code 4: To Create Box Plot

Code 5: To Create Scatter Plot

Conclusion

Artificial Intelligence in Cybersecurity: Pitting Algorithms vs Algorithms

Generate a QR Code using Python

You may also like