Skip to content

Latest commit

 

History

History
191 lines (143 loc) · 7.94 KB

categorical-axes.md

File metadata and controls

191 lines (143 loc) · 7.94 KB
jupyter
jupytext kernelspec language_info plotly
notebook_metadata_filter text_representation
all
extension format_name format_version jupytext_version
.md
markdown
1.2
1.4.2
display_name language name
Python 3
python
python3
codemirror_mode file_extension mimetype name nbconvert_exporter pygments_lexer version
name version
ipython
3
.py
text/x-python
python
python
ipython3
3.7.7
description display_as language layout name order page_type permalink thumbnail
How to use categorical axes in Python with Plotly.
basic
python
base
Categorical Axes
16
example_index
python/categorical-axes/
thumbnail/bar.jpg

This page shows examples of how to configure 2-dimensional Cartesian axes to visualize categorical (i.e. qualitative, nominal or ordinal data as opposed to continuous numerical data). Such axes are a natural fit for bar charts, waterfall charts, funnel charts, heatmaps, violin charts and box plots, but can also be used with scatter plots and line charts. Configuring gridlines, ticks, tick labels and axis titles on logarithmic axes is done the same was as with linear axes.

2-D Cartesian Axis Type and Auto-Detection

The different types of Cartesian axes are configured via the xaxis.type or yaxis.type attribute, which can take on the following values:

The axis type is auto-detected by looking at data from the first trace linked to this axis:

  • First check for multicategory, then date, then category, else default to linear (log is never automatically selected)
  • multicategory is just a shape test: is the array nested?
  • date and category: require more than twice as many distinct date or category strings as distinct numbers or numeric strings in order to choose that axis type.
    • Both of these test an evenly-spaced sample of at most 1000 values
    • Small detail: the category test sorts every value into either number or category, whereas for dates, 2- and 4-digit integers count as both dates and numbers.

Forcing an axis to be categorical

If you pass string values for the x or y parameter, plotly will automatically set the corresponding axis type to category, except if enough of these strings contain numbers as detailed above, in which case the axis is automatically set to linear. It is however possible to force the axis type by setting explicitely xaxis_type to be category.

import plotly.express as px
fig = px.bar(x=[1, 2, 4, 10], y =[8, 6, 11, 5])
fig.update_xaxes(type='category')
fig.show()

Controlling the Category Order with Plotly Express

Plotly Express is the easy-to-use, high-level interface to Plotly, which operates on a variety of types of data and produces easy-to-style figures.

By default, Plotly Express lays out categorical data in the order in which it appears in the underlying data. Every 2-d cartesian Plotly Express function also includes a category_orders keyword argument which can be used to control the order in which categorical axes are drawn, but beyond that can also control the order in which discrete colors appear in the legend, and the order in which facets are laid out.

import plotly.express as px
df = px.data.tips()
fig = px.bar(df, x="day", y="total_bill", color="smoker", barmode="group", facet_col="sex",
             category_orders={"day": ["Thur", "Fri", "Sat", "Sun"],
                              "smoker": ["Yes", "No"],
                              "sex": ["Male", "Female"]})
fig.show()

Automatically Sorting Categories by Name or Total Value

Whether using Plotly Express or not, categories can be sorted alphabetically or by value using the categoryorder attribute:

Set categoryorder to "category ascending" or "category descending" for the alphanumerical order of the category names or "total ascending" or "total descending" for numerical order of values. categoryorder for more information. Note that sorting the bars by a particular trace isn't possible right now - it's only possible to sort by the total values. Of course, you can always sort your data before plotting it if you need more customization.

This example orders the categories alphabetically with categoryorder: 'category ascending'

import plotly.graph_objects as go

x=['b', 'a', 'c', 'd']
fig = go.Figure(go.Bar(x=x, y=[2,5,1,9], name='Montreal'))
fig.add_trace(go.Bar(x=x, y=[1, 4, 9, 16], name='Ottawa'))
fig.add_trace(go.Bar(x=x, y=[6, 8, 4.5, 8], name='Toronto'))

fig.update_layout(barmode='stack')
fig.update_xaxes(categoryorder='category ascending')
fig.show()

This example orders the categories by total value with categoryorder: 'total descending'

import plotly.graph_objects as go

x=['b', 'a', 'c', 'd']
fig = go.Figure(go.Bar(x=x, y=[2,5,1,9], name='Montreal'))
fig.add_trace(go.Bar(x=x, y=[1, 4, 9, 16], name='Ottawa'))
fig.add_trace(go.Bar(x=x, y=[6, 8, 4.5, 8], name='Toronto'))

fig.update_layout(barmode='stack')
fig.update_xaxes(categoryorder='total ascending')
fig.show()

This example shows how to control category order when using plotly.graph_objects by defining categoryorder to "array" to derive the ordering from the attribute categoryarray.

import plotly.graph_objects as go

x=['b', 'a', 'c', 'd']
fig = go.Figure(go.Bar(x=x, y=[2,5,1,9], name='Montreal'))
fig.add_trace(go.Bar(x=x, y=[1, 4, 9, 16], name='Ottawa'))
fig.add_trace(go.Bar(x=x, y=[6, 8, 4.5, 8], name='Toronto'))

fig.update_layout(barmode='stack')
fig.update_xaxes(categoryorder='array', categoryarray= ['d','a','c','b'])
fig.show()

Gridlines, Ticks and Tick Labels

By default, gridlines and ticks are not shown on categorical axes but they can be activated:

import plotly.express as px

fig = px.bar(x=["A","B","C"], y=[1,3,2])
fig.update_xaxes(showgrid=True, ticks="outside")
fig.show()

By default, ticks and gridlines appear on the categories but the tickson attribute can be used to move them to the category boundaries:

import plotly.express as px

fig = px.bar(x=["A","B","C"], y=[1,3,2])
fig.update_xaxes(showgrid=True, ticks="outside", tickson="boundaries")
fig.show()

Multi-categorical Axes

A two-level categorical axis (also known as grouped or hierarchical categories, or sub-categories) can be created by specifying a trace's x or y property as a 2-dimensional lists. The first sublist represents the outer categorical value while the second sublist represents the inner categorical value. This is only possible with plotly.graph_objects at the moment, and not Plotly Express.

Passing in a two-dimensional list as the x or y value of a trace causes the type of the corresponding axis to be set to multicategory.

Here is an example that creates a figure with 2 bar traces with a 2-level categorical x-axis.

import plotly.graph_objects as go

fig = go.Figure()

fig.add_trace(go.Bar(
  x = [['First', 'First', 'Second', 'Second'],
       ["A", "B", "A", "B"]],
  y = [2, 3, 1, 5],
  name = "Adults",
))

fig.add_trace(go.Bar(
  x = [['First', 'First', 'Second', 'Second'],
       ["A", "B", "A", "B"]],
  y = [8, 3, 6, 5],
  name = "Children",
))

fig.update_layout(title_text="Multi-category axis")

fig.show()

Reference

See https://plotly.com/python/reference/layout/xaxis/ for more information and chart attribute options!