1 求滚动\(\rho\)(Vaughan and Dancho 2018)

set.seed(123)
x <- rnorm(100)
set.seed(456)
y <- rnorm(100)
x %>% head()
y %>% head()
t <- 1:100
t %>% head()
test_roll_cor <- 
  data_frame(x = x, y = y, t = t)

rolling_cor <- rollify(.f = ~cor(.x,.y), window = 15)

test_roll_cor %>% 
  mutate(rolling_cor = rolling_cor(x,y))

2 as_tbl_time

weather_time <- nycflights13::weather %>% 
  select(origin, time_hour, temp, precip) %>% 
  as_tbl_time(index = time_hour)

这里使用nycflights13::weather的数据，定义time index为time_hour。

weather_time %>% head

3 collapse_by

weather_time %>% 
  group_by(origin) %>% 
  collapse_by(period = "monthly")

tibbletime collapse_by Evaluation error: invalid 'tz' value. 但是会出现这个bug，已经向作者报错了， collapse_by error invalid timezone · Issue #51 · business-science/tibbletime

4 filter time(Wang 2018)

比原生的dplyr方便。 这里已经指定了time_hour为time index，因此直接给规则。

weather_time %>% 
  filter_time(~ "2013-01-02") %>% 
  arrange(desc(time_hour))