Adding time trend variable in ddply function

Question

I am having difficulties adding a time trend variable within a ddply function,

Here is the simplified code:

id <- c(1,1,1,2,2,2)
price <- c(1,2,3,2,1,0)

df <- data.frame(id, price)

price_trends <- ddply(df, ~id, summarise, 
      mean_price = mean(price), 
      sd_price = sd(price))
price_trends

Now I also want to include a time trend coefficient per id (i.e., price inclining for id 1, declining for id 2), but I am having difficulties to include it in the ddply function from above. I appreciate any help.

Answer 1

You could use:

slope <- function(x) (tail(x,1)-x[1])/length(x)
price_trends <- ddply(df, ~id, summarise, 
                      mean_price = mean(price), 
                      sd_price = sd(price),
                      trend = price %>% slope() %>% round(0))
> price_trends
  id mean_price sd_price trend
1  1          2        1     1
2  2          1        1    -1

Note that you actually get the slope if you remove the last round(0). Further, you can define any coefficient of interest via the function. You can also include the function directly inside ddply like so:

price_trends <- ddply(df, ~id, summarise, 
                      mean_price = mean(price), 
                      sd_price = sd(price),
                      trend = ((tail(price,1)-price[1])/length(price)) %>% round(0)
                      )
price_trends