Plotly – Make your scatterplot wonderful in Python

In this simple tutorial, we’ll see how to create a remarkable scatterplot in Python using one of the best APIs for data science and visualization, of course, I’m talking of Plotly a fantastic API for charts.

Scatterplot

As you’re probably aware, scatterplots allow the comparison of two numeric variables for a set of data. Depending on the trend of the scatter points, we could interpret a correlation between those variables. Let’s jump to code!

Scatterplot in Plotly

In Plotly is easy to create wonderful customized charts, let’s see scatterplots in a few lines of codes

import pandas as pd
import plotly.offline as py
import plotly.graph_objects as go
import numpy as np

np.random.seed(42)

# Create our x and y data
random_x = np.random.randint(1, 101, 100)
random_y = np.random.randint(1, 101, 100)

# In plotly the charts are named traces
trace = go.Scatter(x = random_x, y = random_y, mode='markers')

layout = go.Layout(title='My Simple Scatterplot',
                  xaxis = {'title': 'Random X'},
                  yaxis = {'title': 'Random Y'})
fig = go.Figure(data = trace, layout = layout)
py.plot(fig, filename='scatter-demo')

The previous one is the simplest scatterplot that you can make with Plotly through the module graph_objects. In particular, the class Scatter is employed to reach our objective passing the following, basic, parameters:

X: is a list of numeric values to show along the x-axis
Y: is a list of numeric values to show along the y-axis
Mode: define the modality of the draw, other possible values are ‘lines’, ‘line+markers’ and so on (see my tutorial on how to make a line chart)

You can refer to the documentation to discover the several parameters that the Scatter class has.

Sophisticated Scatter Plot

Now, let’s try to customize our chart playing on those parameters.

import pandas as pd
import numpy as np
import plotly.graph_objects as go
import plotly.offline as py

np.random.seed(42)

# Create our x and y data
random_x = np.random.randint(1, 101, 100)
random_y = np.random.randint(1, 101, 100)

# In plotly the charts are named traces
trace = go.Scatter(x = random_x, y = random_y, mode='markers',
                   selectedpoints=[0, 10, 50, 90], # Points' indexes you want highlight
                   selected=dict( # Define a style for the selected points
                     marker=dict(
                       color = 'rgb(255, 60, 60)', # color
                       size = 24 # in px
                   )),
                   marker=dict( # Define the style of overall points in the chart
                       color = 'rgb(51, 204, 153)', # color
                       size = 12, # in px
                       symbol = 'hexagon' # customized marker symbol
                   ))

layout = go.Layout(title='My Sophisticated Scatterplot',
                  xaxis = {'title': 'Random X'},
                  yaxis = {'title': 'Random Y'})

fig = go.Figure(data = trace, layout = layout)
py.plot(fig, filename='scatter-demo')

It is possible to notice that this chart is a little bit more sophisticated than the first one. Indeed, we added more information to the chart through the parameters of the class Scatter. Just a few comments on the code.

Passing the data

First, as in the previous chart, we passed our data and set markers as the mode

x = random_x, y = random_y, mode='markers'

Selecting the Points

Second, we select some random points of the distribution passing a list of points index to the parameter selected points. Below are details from the documentation:

SelectedPoints: Array containing integer indices of selected points. Has an effect only for traces that support selections

selectedpoints=[0, 10, 50, 90]

Define selected markers’ style

Third, we define the style of the selected points changing the colour (as red) and the size. Below is the parameter description:

Selected: Instance or dict with compatible properties. One of the properties is marker that define the style of the selected points (refer to documentation)

# Define a style for the selected points
selected=dict(marker=dict(color = 'rgb(255, 60, 60)', # color
                       	  size = 24))

The general style

Finally, we define a style for the overall points in the chart choosing the symbol (as hexagon), the size (as 12px) and the colour (as green), in particular for the colours we used the RGB scale

# Define the style of overall points in the chart
marker=dict(color = 'rgb(51, 204, 153)', # color
            size = 12, # in px
            symbol = 'hexagon') # customized marker symbol

Note: you can define markers style passing the class Marker or a dictionary-like data structure. All the possible configurations can be discovered in the documentation.