I have a time series of cost values and I'd like to get the sum of the costs for a range of selected dates. The interval selection is working perfectly, but how do I access the selected dates?
brush = alt.selection(type='interval', encodings=['x'])
chart = alt.Chart(forecast).mark_bar().encode(
x='dates:T',
y='costs:Q',
color='type:N',
).add_selection(
brush
)
You can use a text mark along with a filter transform to display the sum of the values within the selection. For example:
import altair as alt
import pandas as pd
import numpy as np
data = pd.DataFrame({
"dates": pd.date_range('2019-04-01', freq='M', periods=20),
"costs": np.linspace(1000, 5000, 20),
})
brush = alt.selection(type='interval', encodings=['x'])
chart = alt.Chart(data).mark_bar().encode(
x='dates:T',
y='costs:Q',
).add_selection(
brush
)
text = alt.Chart(data).transform_filter(brush).mark_text(
align='left',
baseline='top',
).encode(
x=alt.value(5),
y=alt.value(5),
text=alt.Text('sum(costs):Q', format='.1f'),
)
chart + text
Edit: if you want to display the selection bounds in the text, it's not entirely straightforward, but you can do it using a calculate transform along with an appropriate vega expression string which refers to the selection by name.
For example:
brush = alt.selection(type='interval', encodings=['x'], name='sel')
chart = alt.Chart(data).mark_bar().encode(
x='dates:T',
y='costs:Q',
).add_selection(
brush
)
text = alt.Chart(data).transform_filter(
brush
).transform_aggregate(
total='sum(costs)'
).transform_calculate(
date_range="sel.dates ? monthAbbrevFormat(month(sel.dates[0])) + ' to ' + monthAbbrevFormat(month(sel.dates[1])) : 'all'",
text="'Total for ' + datum.date_range + ': ' + format(datum.total, '.0f')"
).mark_text(
align='left',
baseline='top',
).encode(
x=alt.value(5),
y=alt.value(5),
text=alt.Text('text:N'),
)
chart + text
