Retrieve most frequent value for each couple of values

I'm working with a df having 3 columns: 'Latitude', 'Longitude' and 'street_type'.

The 'street_type' column has multiple values like 'rectilinear', 'roundabout' exc.

I would like to retrieve the most frequent value for each couple of values 'Latitude' and 'Longitude'.

I tried using the group by expression like this one:

df=df.group_by(['Latitude','Longitude'])[['steet_type']].sum().reset_index()

It returns me the df with the group by but I would like to have the most frequent value with its count and the coordinate for that value.

The street_type column has only strings si I used the get_dummies method to have only numbers. Was it a good practice or not for the groupby() and a sum()?

Thank you in advance!

data = {'Latitude'=[-41.794,-41.733,-41.335,-41.794],'Longitude'=[-12.344,-12.456,-12.929,-12.344],'street_type'={'rectilinear','roundabout','intersection','rectilinear']}

df = pd.DataFrame(data)
df = pd.get_dummies(df, columns=['street_type'])
df = df.groupby(['Latitude','Longitude'])[['street_type_rectilinear','street_type_roundabout','street_type_intersection']].sum().reset_index()

The output from groupby snd sum()

same result for agg(pd.Series.mode) same result for agg(lambda x: x.mode().iat[0]).reset_index()

The desidered output is:

EDITED THE QUESTION WITH SAMPLE CODE