如何对某些观察结果运行函数？

Question

I am trying to calculate the number of seats that parties won per district in a given election based on the number of votes they received in the district.

There is a function in R that will do this for each region:

seats_ha(party, votes, seats, method="dhondt")

The first argument provides a vector of party list names, the second argument provides a vector of the vote totals that each party won in a district, seats = the number of seats in a given district and the method is the electoral formula used to translate votes into seats. What I've been able to do is to calculate this by individually subsetting the data by a single region in an electoral year. My problem is that I have ~27 regions over 3 electoral years.

So my data look like this:

year   region dist_seat  party_name party_vote reg_id cong_id
2016-2021 AMAZONAS   2          UPP     0            1       3
2016-2021 AMAZONAS   2          FP      51067        1       3
2016-2021 AMAZONAS   2          AP      11992        1       3
2016-2021   ANCASH   5          FE       4534        2       3
2016-2021   ANCASH   5          UPP     0            2       3

I would like to be able to run the function for each region in each year.

Answer 1

Consider by , the object-oriented wrapper to tapply which slices a data frame by one or more columns to run needed operations. The input parameter to by's FUN is always a subsetted data frame and output will always be a list of whatever function returns, here being the return of seat_ha .

You can even add a new column to subsetted data frame and then do.call + rbind results for a single data frame. Below tryCatch ensures new column is always populated: actual result of seats_ha and NA if it encounters an error.

BUILD LIST OF SUBSETTED DFs
df_list <- by(mydata, mydata[,c("year", "region")], FUN=function(sub) {
                 # ADD NEW COLUMN TO sub DF
                 sub$calc_seat <- tryCatch(with(sub, seats_ha(party_name, party_vote, 
                                                              dist_seat, method="dhondt")), 
                                           error = function(e) NA)
                 return(sub)
             })

# ROW BIND ALL DFs
final_df <- do.call(rbind, df_list)