Sum group by label in Pandas

Asked Feb 15 '19 at 18:13

Active Feb 15 '19 at 18:52

Viewed 23 times

I have a set of data with various stock symbols and would like to know how many bullish and bearish signals there were for each symbol for each day.

How would I do this with Pandas? Thanks

Input:      
Date    Symbol  Sentiment
5/2/19  GE      Bullish
5/2/19  GE      Bullish
5/2/19  GE      Bearish
5/2/19  GE      Bearish
6/2/19  GE      Bullish
6/2/19  GE      Bullish
6/2/19  GE      Bullish
6/2/19  GE      Bullish
6/2/19  GE      Bullish
6/2/19  GE      Bearish
6/2/19  GE      Bearish
6/2/19  GE      Bearish



Output:         
Date    Symbol  Sentiment Bullish   Sentiment Bearish
5/2/19  GE      2                   2
6/2/19  GE      5                   3

edited Feb 15 '19 at 18:52

smci

29,564
18
109
144

asked Feb 15 '19 at 18:13

nicholas.reichel

2,190
6
17
27

1

can you post the data as text please. :) – anky Feb 15 '19 at 18:14
1

Specifically, see question 9, using `.pivot_table` with `aggfunc='size`, i.e. `df.pivot_table(index=['Date', 'Symbol'], columns='Sentiment', aggfunc='size')` – ALollz Feb 15 '19 at 18:20
Ok I think I got it. Thanks. – nicholas.reichel Feb 15 '19 at 18:25
1

`crosstab` is really well suited for this too: `pd.crosstab(index=[df.Date, df.Symbol], columns=df.Sentiment)`. Also possible with `.groupby`. – ALollz Feb 15 '19 at 18:29
Wonderful. I was just about to make a Date, Symbol index but this works too. Thanks – nicholas.reichel Feb 15 '19 at 18:33

Sum group by label in Pandas

0 Answers0