创建在 X 轴上按周分组日期的 Altair 图表

Question

DATE,AMOUNT
2022-04-05,100
2022-04-06,10
2022-04-07,90
2022-04-08,75
2022-04-12,32
2022-04-13,400
2022-04-14,28
2022-04-15,50

With a dataset like this, how can I create a bar chart grouped by week so the X axis shows only two bars, april 03 - april 09 and april 11 - april 17 ? (Taking into account that the week starts on sundays, even tho there is no data for dates like april 04 )

Answer 1

You can use the time units in VegaLite to group observations. week() will return the week number, but I am not sure there is a way to format the label the way you want without doing it manually via labelExpr as I have done below.

import pandas as pd
import altair as alt

df = pd.read_clipboard(sep=',')

alt.Chart(df).mark_bar().encode(
    x='AMOUNT',
    y=alt.Y(
        'week(DATE):O', axis=alt.Axis(
            labelExpr="datum.label == 'W14' ? 'Apr 04 - Apr 10' : 'Apr 11 - Apr 17'")
   )
)

You could also compute the labels via pandas first, which is more automatic than the above:

df['DATE'] = pd.to_datetime(df['DATE'])
df_by_week = df.resample('W', on ='DATE').sum().reset_index()
df_by_week['date_label'] = df_by_week['DATE'].apply(
    lambda x: f'{(x - pd.Timedelta(days=6)).strftime("%b %d")} - {x.strftime("%B %d")}'
)
alt.Chart(df_by_week).mark_bar().encode(
    x='AMOUNT',
    y='date_label'
)