ES last five hours of the day ATR

[SunTrader]

Easy to do this with Databento:
....

Here's what you're looking for over the last month:
2022-12-19 35.50
2022-12-20 22.50
.....

Not last 5 hours .... "for each of the last five hours".

 

[Databento]

Not last 5 hours .... "for each of the last five hours".

Oh sorry, misread. That's pretty easy too, just switch out the last 2 lines:

df['range'] = df['high'] - df['low']
df['atr14'] = df['range'].rolling(14).mean()

# Only last 5 hours of each day
df = df.between_time('12:00', '17:00')

Result:

ts_event
2022-12-19 12:00:00-05:00    11.089286
2022-12-19 13:00:00-05:00    12.303571
2022-12-19 14:00:00-05:00    12.857143
2022-12-19 15:00:00-05:00    13.839286
2022-12-19 16:00:00-05:00    13.910714
                               ...
2023-01-17 12:00:00-05:00     9.678571
2023-01-17 13:00:00-05:00    10.089286
2023-01-17 14:00:00-05:00    10.732143
2023-01-17 15:00:00-05:00    11.339286
2023-01-17 16:00:00-05:00    11.214286