Bokeh – Quick Guide ”; Previous Next Bokeh – Introduction Bokeh is a data visualization library for Python. Unlike Matplotlib and Seaborn, they are also Python packages for data visualization, Bokeh renders its plots using HTML and JavaScript. Hence, it proves to be extremely useful for developing web based dashboards. The Bokeh project is sponsored by NumFocus https://numfocus.org/. NumFocus also supports PyData, an educational program, involved in development of other important tools such as NumPy, Pandas and more. Bokeh can easily connect with these tools and produce interactive plots, dashboards and data applications. Features Bokeh primarily converts the data source into a JSON file which is used as input for BokehJS, a JavaScript library, which in turn is written in TypeScript and renders the visualizations in modern browsers. Some of the important features of Bokeh are as follows − Flexibility Bokeh is useful for common plotting requirements as well as custom and complex use-cases. Productivity Bokeh can easily interact with other popular Pydata tools such as Pandas and Jupyter notebook. Interactivity This is an important advantage of Bokeh over Matplotlib and Seaborn, both produce static plots. Bokeh creates interactive plots that change when the user interacts with them. You can give your audience a wide range of options and tools for inferring and looking at data from various angles so that user can perform “what if” analysis. Powerful By adding custom JavaScript, it is possible to generate visualizations for specialised use-cases. Sharable Plots can be embedded in output of Flask or Django enabled web applications. They can also be rendered in Jupyter notebooks. Open source Bokeh is an open source project. It is distributed under Berkeley Source Distribution (BSD) license. Its source code is available on https://github.com/bokeh/bokeh. Bokeh – Environment Setup Bokeh can be installed on CPython versions 2.7 and 3.5+ only both with Standard distribution and Anaconda distribution. Current version of Bokeh at the time of writing this tutorial is ver. 1.3.4. Bokeh package has the following dependencies − jinja2 >= 2.7 numpy >= 1.7.1 packaging >= 16.8 pillow >= 4.0 python-dateutil >= 2.1 pyyaml >= 3.10 six >= 1.5.2 tornado >= 4.3 Generally, above packages are installed automatically when Bokeh is installed using Python’s built-in Package manager PIP as shown below − pip3 install bokeh If you are using Anaconda distribution, use conda package manager as follows − conda install bokeh In addition to the above dependencies, you may require additional packages such as pandas, psutil, etc., for specific purposes. To verify if Bokeh has been successfully installed, import bokeh package in Python terminal and check its version − >>> import bokeh >>> bokeh.__version__ ”1.3.4” Bokeh – Getting Started Creating a simple line plot between two numpy arrays is very simple. To begin with, import following functions from bokeh.plotting modules − from bokeh.plotting import figure, output_file, show The figure() function creates a new figure for plotting. The output_file() function is used to specify a HTML file to store output. The show() function displays the Bokeh figure in browser on in notebook. Next, set up two numpy arrays where second array is sine value of first. import numpy as np import math x = np.arange(0, math.pi*2, 0.05) y = np.sin(x) To obtain a Bokeh Figure object, specify the title and x and y axis labels as below − p = figure(title = “sine wave example”, x_axis_label = ”x”, y_axis_label = ”y”) The Figure object contains a line() method that adds a line glyph to the figure. It needs data series for x and y axes. p.line(x, y, legend = “sine”, line_width = 2) Finally, set the output file and call show() function. output_file(“sine.html”) show(p) This will render the line plot in ‘sine.html’ and will be displayed in browser. Complete code and its output is as follows from bokeh.plotting import figure, output_file, show import numpy as np import math x = np.arange(0, math.pi*2, 0.05) y = np.sin(x) output_file(“sine.html”) p = figure(title = “sine wave example”, x_axis_label = ”x”, y_axis_label = ”y”) p.line(x, y, legend = “sine”, line_width = 2) show(p) Output on browser Bokeh – Jupyter Notebook Displaying Bokeh figure in Jupyter notebook is very similar to the above. The only change you need to make is to import output_notebook instead of output_file from bokeh.plotting module. from bokeh.plotting import figure, output_notebook, show Call to output_notebook() function sets Jupyter notebook’s output cell as the destination for show() function as shown below − output_notebook() show(p) Enter the code in a notebook cell and run it. The sine wave will be displayed inside the notebook. Bokeh – Basic Concepts Bokeh package offers two interfaces using which various plotting operations can be performed. bokeh.models This module is a low level interface. It provides great deal of flexibility to the application developer in developing visualizations. A Bokeh plot results in an object containing visual and data aspects of a scene which is used by BokehJS library. The low-level objects that comprise a Bokeh scene graph are called Models. bokeh.plotting This is a higher level interface that has functionality for composing visual glyphs. This module contains definition of Figure class. It actually is a subclass of plot class defined in bokeh.models module. Figure class simplifies plot creation. It contains various methods to draw different vectorized graphical glyphs. Glyphs are the building blocks of Bokeh plot such as lines, circles, rectangles, and other shapes. bokeh.application Bokeh package Application class which is a lightweight factory for creating Bokeh Documents. A Document is a container for Bokeh Models to be reflected to the client side BokehJS library. bokeh.server It provides customizable Bokeh Server Tornadocore application. Server is used to share and publish interactive plots and apps to an audience of your choice. Bokeh – Plots with Glyphs Any plot is usually made up of one or many geometrical shapes such as line, circle, rectangle, etc. These shapes have visual information about the corresponding set of data. In Bokeh terminology, these geometrical shapes are called gylphs. Bokeh plots constructed using bokeh.plotting interface use a default set