I have a data frame with products and their variances that I want to plot as a histogram:

import pandas as pd
data = {'product_name': ['link', 'zelda', 'impa', 'rauru', 'saria', 'darunia'], 'variance': [0.95, -0.85, 0.3, 0.2, 0.03, 0.02]}
df = pd.DataFrame(data)

I can use plotly to create a histogram:

px.histogram(df, 'variance')

But I want to create five bins, with custom start and ends (like in the graph below) rather than what plotly defaults to. This seems weirdly difficult to do using plotly express - is there a way to customize the bin sizes at all?

Thanks

I have a data frame with products and their variances that I want to plot as a histogram:

import pandas as pd
data = {'product_name': ['link', 'zelda', 'impa', 'rauru', 'saria', 'darunia'], 'variance': [0.95, -0.85, 0.3, 0.2, 0.03, 0.02]}
df = pd.DataFrame(data)

I can use plotly to create a histogram:

px.histogram(df, 'variance')

But I want to create five bins, with custom start and ends (like in the graph below) rather than what plotly defaults to. This seems weirdly difficult to do using plotly express - is there a way to customize the bin sizes at all?

Thanks

• python
• pandas
• plotly
• histogram

1 Answer 1

A way to deal with this is to use pandas' cut method to bin the variance columns into user-defined categories:

bins = [-1, -0.25, -.05, 0.05, 0.25, 1]
names = ['-25%+', '-5 to -25%', '-5 to 5%', '5 to 25%', '25%+']
df['MyCategories'] = pd.cut(df['variance'], bins, labels=names)

You may then create your histogram as follows:

px.histogram(df, 'MyCategories', category_orders=dict(variance = names))

...which returns: