Plotly bar plot colored by week?

📊 Plotly Python
1

I am trying to make a what is a simple plot in excel but can’t figure out how to do it in plotly.

image
image856×537 16.8 KB

I’ve tried

px.bar(df, x=['A','B','D','E'], color='date')

but gives a key error.

2

hi @benromarowski

If your data was in a long format, you could use this:

import pandas as pd
import plotly.express as px

categories = ['A','A','B','B','D','D','E','E']
values = [0,20,0,20,80,40,20,0]
dates = ['2022-12-05','2022-12-08','2022-12-05','2022-12-08','2022-12-05','2022-12-08','2022-12-05','2022-12-08']

df = pd.DataFrame({'date':dates,
                   'cat':categories,
                   'val':values
                   })
print (df)

fig = px.bar(df, x='cat', y='val',  color='date', barmode='group')

fig.show()

To convert your data from wide to long you can do:

df_melted = pd.melt(df,
                id_vars=['date'],
                var_name='cat',
                value_name='val')

print (df_melted)

So the full code with the current data would be:

import pandas as pd
import plotly.express as px

Acol = [0,20]
Bcol = [0,20]
Dcol = [80,40]
Ecol = [20,0]
dates = ['2022-12-05','2022-12-08']


df = pd.DataFrame({'date':dates,
                   'A':Acol,
                   'B':Bcol,
                   'D':Dcol,
                   'E':Ecol
                   })
print(df)
df_melted = pd.melt(df,
                id_vars=['date'],
                var_name='cat',
                value_name='val')

print (df_melted)

fig = px.bar(df_melted, x='cat', y='val',  color='date', barmode='group')

fig.show()

Resulting in:

image
image1920×1080 55.1 KB

2 Likes
3

Hi Adam, thanks for your answer, seems like melting the DF is the way to go.
I still get a key error for the Timestamp as if they were not recognized by plotly, do you know why?

4

I fixed this by parsing the date to str with .astype()

2 Likes