How Do I Aggregate Recorded Daily Incidents Into Monthly Counts


I have a large pandas dataframe of daily crime incidents. One of my columns (‘date’) is datatype datetime64[ns].

Take, for example, the dataframe below:

ID Date Crime
987AL4 1991-08-15 Robbery
987AL4 1991-08-16 Asasult
124576IL 1991-09-21 Breaking and Entering
4689CA 2016-08-17 Sexaul Assault

What I would like to do is create a new dataframe with monthly counts of incidents for each year and each ID.
So, ideally, the resulting dataframe might look something like:

ID Month Count
987AL4 1991-08 200,870
124576IL 1991-08 190,000
45678CA 2016-08 60,000

I’ve tried various groupby and count methods, but I’ve had no luck. I would greatly appreciate any help.

Asked By: TheMaffGuy



You can convert your DatetimeIndex as PeriodIndex and use value_counts:

out = (df.value_counts(['ID', df['Date'].dt.to_period('M').rename('Month')])

# Output
         ID    Month  Count
0    987AL4  1991-08      2
1  124576IL  1991-09      1
2    4689CA  2016-08      1
Answered By: Corralien