How to group one column and rank values by grouped totals of another column

Question

Situation

I have a dataframe of tweets pulled from the Twitter API. Each tweet has an author_id parameter, and a retweets and an engagements parameter.

I'd like to group the tweets by author_id, as some authors have posted more than one tweet, and then rank them by the total number of engagements from the sum of each author's tweets.

Example Dataframe

index	text	author_id	retweets	engagement
2022-03-20	Max Muncy is great!	100	3	2
2022-03-21	The worst Dodger is Max muncy.	101	1	8
2022-03-24	Max Muncy was great, but not so much now.	101	1	4
2022-03-28	What a fantastic guy, that Max muncy.	102	4	5
2022-03-31	Max Muncy, my vote for MVP.	103	8	17

What I tried

df01a.groupby(['author_id']).sum()

That returned this:

author_id	text	retweets	engagement
100	Max Muncy is great!	3	2
101	The worst Dodger is Max muncy.Max Muncy was great, but not so much now.	11	84
102	What a fantastic guy, that Max muncy.	4	5
103	Max Muncy, my vote for MVP.	8	17

Assessment

The first thing I notice is that the retweet and engagement figures are concatenating instead of summing, and I've not been able to sort out why, as the sum method seems pretty straight-forward.

Thank you for any input on what I might be doing wrong here.

UPDATE - 4/5

Just noticed the note on datatypes so posting that here:

text          object
author_id     object
retweets      object
engagement    object
dtype: object

Answer 1

First make sure the engagement column is numerical and not a string:

df['engagement'] = pd.to_numeric(df['engagement'])

From what you are saying you do not need to use groupby here. Simply sort by author_id and engagment.

df.sort_values(['author_id', 'engagement'], ascending=[True, True])