Plotly Strip plot - color by continuous scale

Question

I would like to use plotly.express.strip to show the distribution of my data points. The catch is that I would also like to color each data point according to its value. As an example, taken from Plotly's documentation:

import plotly.express as px

df = px.data.tips()
fig = px.strip(df, x="total_bill", y="day")
fig.show()

This shows: But I would like the points to be colored by the "total_bill" amount using a continuous color scale.

From my research I haven't found a way to do this using this specific plot (ie I can't use a px.scatter).

Is there any way to do this? Thanks

Answer 1

You need a jitter. I applied a pseudo-jitter function to the y-axis to create a scatter plot. The jitter function was inspired by the contents of this page . It may differ from the original graphing logic. At first, I wrote the graph by extracting the days of the week in a loop process, but since the color scale duplicates, I described all the days of the week and added the color scale only for Saturday since Saturday contains the largest value.

import plotly.graph_objects as go
import plotly.express as px
import numpy as np

df = px.data.tips()
N = len(df)

fig = go.Figure()

dfs = df.query('day == "Sun"')
fig.add_trace(go.Scatter(
    x=dfs['total_bill'], 
    y=0 + np.random.rand(N) * 0.2,
    mode='markers',
    marker=dict(
        size=9,
        color=dfs['total_bill'],
    ),
    name='Sun',
))

dfst = df.query('day == "Sat"')
fig.add_trace(go.Scatter(
    x=dfst['total_bill'], 
    y=1 + np.random.rand(N) * 0.2,
    mode='markers',
    marker=dict(
        size=9,
        color=dfst['total_bill'],
        colorbar=dict(
            title='total_bill',
        ),
        colorscale='Plasma'
    ),
    name='Sat',
))

dfth = df.query('day == "Thur"')
fig.add_trace(go.Scatter(
    x=dfth['total_bill'], 
    y=2 + np.random.rand(N) * 0.2,
    mode='markers',
    marker=dict(
        size=9,
        color=dfth['total_bill'],
    ),
    name='Thur',
))

dff = df.query('day == "Fri"')
fig.add_trace(go.Scatter(
    x=dff['total_bill'], 
    y=3 + np.random.rand(N) * 0.2,
    mode='markers',
    marker=dict(
        size=9,
        color=dff['total_bill'],
    ),
    name='Fri',
))

fig.update_layout(showlegend=False, coloraxis_showscale=True)
fig.update_yaxes(tickvals=[0,1,2,3], ticktext=['Sun','Sat','Thur','Fri'])
fig.show() 

Answer 2

You can use Plotly or Seaborn

Plotly

I am using grays here (white to black), but you can choose what makes sense. Check the RGB values... n_colors will break up the colors into groups. Used 4 as there are 4 unique days here

Code:

import plotly.express as px
from plotly.colors import n_colors

df = px.data.tips()
fig = px.strip(df, x="total_bill", y="day", color="day", color_discrete_sequence = n_colors('rgb(0, 0, 0)', 'rgb(255, 255, 255)', 4, colortype = 'rgb'))

Output

Seaborn

It is actually easier in seaborn as you can use palette and using jitter=True , you should be good to go

Code:

import seaborn as sns
sns.stripplot(data = df, x='total_bill', y = 'day', palette = 'gray', jitter=True)

Output

Option 3 - continuous X-axis If you need the gradient on the X-axis, you need add hue='total_bill' to the sns plot. Code below... I have removed the legend as it will show all colors, which you probably don't need

import seaborn as sns
sns.stripplot(data = df, x='total_bill', y = 'day', hue = 'total_bill', palette = 'gray', jitter=True)
plt.legend([],[], frameon=False)

Output