weighted_average function in calculations-latest_polls_weights.R

[mhoehle]

You may also want to consider using the summarise function instead of mutating an entire column and then extract the first element, which seems a bit clumsy. This also goes for the weighted_error function. For example:

weighted_average <- function(df_in,party_in,date_in){
  sz_avr <- df_in %>%
    tbl_df %>%
    filter(datum <= date_in) %>%
    filter(partei == party_in) %>%
    group_by(institut) %>%
    filter(datum == max(datum)) %>%
    ungroup() %>%
    summarise(roll_avr = sum(befragte * anteil) / sum(befragte)) %>% as.numeric
  sz_avr
}