Plotly-Python: How To Make A Gapped Y Axis?

Courtesy to the original question for Plotly-R. The following will focus on Python. Is it possible to create a plotly bar chart, e.g. any chart from the following website: plotly.

Solution 1:

To my knowledge, plotly hasn't got any built-in functionality to do this. But it's still possible to make a figure that matches your image using subplots if you:

1. use make_subplots(rows=2, cols=1, vertical_spacing = <low>),
2. add the same traces to figure positions [1, 1] and [2, 1],
3. remove x-axis labels for [1, 1], and
4. adjust the y axes for figure positions [1, 1] and [2, 1] to respectively start and end with your desired cutoff values in a defined interval.

Plot:

Complete code:

# imports
import plotly.graph_objects as go
import plotly.express as px
import pandas as pd
from plotly.subplots import make_subplots

# data
df = pd.DataFrame({'years': [1995, 1996, 1997, 1998, 1999, 2000,
                             2001, 2002, 2003, 2004, 2005, 2006,
                             2007, 2008, 2009, 2010, 2011, 2012],
                  'China': [219, 146, 112, 127, 124, 180, 236,
                            207, 236, 263,350, 430, 474, 1526,
                            488, 537, 500, 439],
                  'Rest of world': [16, 13, 10, 11, 28, 37,
                                        43, 55, 56, 88, 105, 156, 270,
                                        299, 340, 403, 549, 1499]})
df.set_index('years', inplace = True)

# colors and cut-offs
colors = px.colors.qualitative.Plotly
cut_interval = [600, 1400]

# subplot setup
fig = make_subplots(rows=2, cols=1, vertical_spacing = 0.04)
fig.update_layout(title = "USA plastic scrap exports (...with some made-up values)")

# Traces for [2, 1]
# marker_color=colors[i] ensures that categories follow the same color cycle
for i, col in enumerate(df.columns):
    fig.add_trace(go.Bar(x=df.index,
                    y=df[col],
                    name=col,
                    marker_color=colors[i],
                    legendgroup = col,
                    ), row=2, col=1)

# Traces for [1, 1]
# Notice that showlegend = False.
# Since legendgroup = col the interactivity is
# taken care of in the previous for-loop.
for i, col in enumerate(df.columns):
    fig.add_trace(go.Bar(x=df.index,
                    y=df[col],
                    name=col,
                    marker_color=colors[i],
                    legendgroup = col,
                    showlegend = False,
                    ), row=1, col=1)

# Some aesthetical adjustments to layout
fig.update_yaxes(range=[cut_interval[1], max(df.max()*1.1)], row=1, col=1)
fig.update_xaxes(visible=False, row=1, col=1)
fig.update_yaxes(range=[0, cut_interval[0]], row=2, col=1)

fig.show()