使用dplyr统计一周各日不同用户类型的平均骑行人数

我的数据样例如下：

# A tibble: 4,722,462 x 5
   started_at          member_casual weekday   ride_length month 
   <dttm>              <chr>         <fct>           <dbl> <fct> 
 1 2020-08-20 18:08:14 member        Thursday        0.160 August
 2 2020-08-27 18:46:04 casual        Thursday        1.15  August
 3 2020-08-26 19:44:14 casual        Wednesday       2.15  August
 4 2020-08-27 12:05:41 casual        Thursday        0.801 August
 5 2020-08-27 16:49:02 casual        Thursday        0.180 August
 6 2020-08-27 17:26:23 casual        Thursday        0.691 August
 7 2020-08-26 20:14:02 casual        Wednesday       0.333 August
 8 2020-08-26 21:59:50 casual        Wednesday       0.212 August
 9 2020-08-26 19:17:42 casual        Wednesday       0.242 August
10 2020-08-27 15:13:57 casual        Thursday        0.467 August
# ... with 4,722,452 more rows

我需要按weekday和member_casual字段分组汇总，得到一周内每一天对应不同用户类型的平均骑行人数，例如「周一」和casual（散客）对应行的计算规则为：数据中周一与散客同时出现的总次数 ÷ 给定时间范围内实际的周一总天数。目前我已写出如下接近需求的实现代码：

# 计算给定时间范围内的总周数
weeks_ <- as.numeric(difftime(max(df2$started_at),min(df2$started_at),units="weeks"))
# 假设时间范围均为完整周
df2 %>% group_by(weekday,member_casual)%>% summarise("Average Riders"=(n()/weeks_))

由于时间范围足够大，该输出虽不十分精确但可满足准确度要求，输出结果示例如下：

weekday   member_casual `Average Riders`
   <fct>     <chr>                    <dbl>
 1 Monday    casual                   4404.
 2 Monday    member                   6688.
 3 Tuesday   casual                   4279.
 4 Tuesday   member                   7289.
 5 Wednesday casual                   4434.
 6 Wednesday member                   7648.
 7 Thursday  casual                   4447.
 8 Thursday  member                   7285.
 9 Friday    casual                   5807.
10 Friday    member                   7452.
11 Saturday  casual                   9366.
12 Saturday  member                   7612.
13 Sunday    casual                   7527.
14 Sunday    member                   6331.

内容的提问来源于stack exchange，提问作者Spinach55