In this post, we will examine additional ways to modify a Bokeh data visualization. The data points in a Bokeh visualization are referred to as glyphs. Glyphs can take various shapes, colors, etc, and we will learn how to modify glyphs specifically for scatter plots in this post.

Libraries

We will need several libraries to work with glyphs in a scatterplot. They are listed below.

from pydataset import data
from bokeh.plotting import figure
from bokeh.io import output_file, show

The first library is for pulling the data we need. The next two libraries are basic libraries from Bokeh for generating our scatterplot.

Data Preparation

Data preparation is simple in this example. All we have to do is use the data() function to load the “Duncan” dataset into an object we call “df.” After this, we print out the first few rows of this dataset using the head() method.

df=data('Duncan')
df.head()

The Duncan dataset contains various occupations that are scored on type, income, education, and prestige. We will be focused primarily on education and prestige in this post.

Default Scatterplot

The code below provides a default scatterplot without any modifications. The purpose of providing this is to serve as a comparison to what we will do when we modify the glyphs in the subsequent examples. Below is the code, output, and explanation.

#Fig Setup
fig = figure(x_axis_label="Prestige", y_axis_label="Education", title="Prestige vs Education")

# Add glyphs
fig.circle(x="prestige", y="education", source=df)

output_file(filename="pres_vs_inc.html")
show(fig)

The first line of code sets up the x and y axis using the figure() function in an object we created called “fig”. The second line allows us to add the actual data points using the .circle() method on the “fig” object. The last two lines allow us to display our plot. output_file() function allows us to save our work and the show() function displays the plot.

Modified Glyphs

Our first modification will involve changes to the appearance of the glyphs in the scatterplot. The first line is the same but the second line includes changes to the size, color, and transparency of the glyphs. Below is the code, output, and explanation.

fig = figure(x_axis_label="Prestige", y_axis_label="Education", title="Prestige vs Education")

# Add glyphs
fig.circle(x="prestige", y="education", source=df, size=16, fill_color="yellow", fill_alpha=0.2)

output_file(filename="pres_vs_inc.html")
show(fig)

In the second line of code, we changed the size to 16 using the size argument and the fill of the dots to yellow using the fill_color argument. We also adjusted the transparency using the fill_alpha argument.

Multiple Glyphs

The example below includes the modification of different glyphs based on a categorical variable. A tooltip is also included to show how various tools can be combined when employing Bokeh. The code, output, and explanation are below.

#Subset Data
source_wc = df[df["type"]=="wc"]
source_prof = df[df["type"]=="prof"]

TOOLTIPS=[('Education', '@education'), ('Prestige', '@prestige')]
fig = figure(x_axis_label="Education", y_axis_label="Prestige", tooltips = TOOLTIPS)
wc_glyphs = fig.circle(x="education", y="prestige", source=source_wc, legend_label="WC", fill_color="blue",fill_alpha=0.2)
prof_glyphs = fig.circle(x="education", y="prestige", source=source_prof, legend_label="Prof", fill_color="green", fill_alpha=0.6)

output_file(filename="update.html")
show(fig)

Here is what we did

1. We subsetted the data into two different objects (source_ws & source_prof) based on the type variable.
2. We created our tooltip which allows us to see the education and prestige of individual data points on the graph.
3. We make the figure or x and y axis for our plot.
4. We use the .circle() method twice. Once for job type “WC” and once for job type “prof”. We set the colors, fills, and transparency of each subset of data. Notice that we created a legend as well. Remember that we named each object in this step wc_glyph and prof_glyph
5. The last two lines of code create a file for the visualization and display it.

Updating Glyphs

We will now learn how to update our code rather than recreating it from scratch. We will update the size of the glyphs and change their color from the prior example.

# Update glyph size
wc_glyphs.glyph.size = 20
prof_glyphs.glyph.size = 10

# Update glyph fill_color
wc_glyphs.glyph.fill_color = "red"
prof_glyphs.glyph.fill_color = "yellow"

output_file(filename="update.html")
show(fig)

We begin by using the objects we created in the last example wc_glyph and prof_glyph, These two objects contain all the information for creating the figure and data points. To update the glyph size we will use the .glyph.size argument and set it to two different values for each of the two objects that were created. We repeat this process for the fill_color.

You can compare the last two visualizations and see the differences for yourself.

Conclusion

Being able to customize glyphs is another powerful feature of Bokeh. With these tools, you can modify your visualizations with ease to communicate with your audience.