r - dplyr: counting the frequency of unique values in one variable for each unique value of another variable in the same data frame

Question

So here's a sample of some of the rows from my dataframe:

> data[1:25, c("TR_DATE", "TR_TYPE...")]
  TR_DATE TR_TYPE...
1  2016-03-01          4
2  2016-03-01          4
3  2016-03-01          5
4  2016-03-01          4
5  2016-03-01          1
6  2016-03-01          7
7  2016-03-01          4
8  2016-03-01          4
9  2016-03-01          24
10 2016-03-01          23
11 2016-03-01          4
12 2016-03-02          4
13 2016-03-02          1
14 2016-03-02          1
15 2016-03-02          4
16 2016-03-02          4
17 2016-03-02          14
18 2016-03-02          4
19 2016-03-02          4
20 2016-03-03          4
21 2016-03-03          1
22 2016-03-03          4
23 2016-03-03          23
24 2016-03-03          1
25 2016-03-03          4

What I'd like to do exactly is rearrange in such a way that for every unique day, I get the number of unique transaction types and the frequency of each transaction type

Here's the code that I tried:

data %>%
group_by(TR_DATE) %>%
summarise(trancount = n(), trantype = n_distinct(TR_TYPE...))

which gave me part of the result that I wanted:

# A tibble: 68 x 3
  TR_DATE trancount trantype
   <date>     <int>    <int>
 1 2016-03-01      5816        6
 2 2016-03-02      5637        3
 3 2016-03-03      4818        3
 4 2016-03-04      5070        8
 5 2016-03-05         4        2
 6 2016-03-08      6707        5
 7 2016-03-09      5228        5
 8 2016-03-10      4722        6
 9 2016-03-11      4469        8
10 2016-03-12         1        1
# ... with 58 more rows

so trantype tells me the number of unique transaction types that happened on a particular day, but I'd like to know the frequency of each of these unique transaction types. What would be the best way to go around doing this? I tried looking around and found similar questions but was unable to modify the solutions to my requirement. I'm fairly new to R and would really appreciate some help. Thanks.

Answer 1

You should group by both variables:

data %>%
group_by(TR_DATE, TR_TYPE...) %>%
summarise(trancount = n(), trantype = n_distinct(TR_TYPE...))