Computing a Lag on a Table in R

Question

I have a data table, my_data, with 3 variables, symbol, date, stock price and 141851 observations. The idea is to compute a Lag of 1 day on the stock price, taking care to group by symbol first, otherwise I will have lags computed using values from the previous stock in the data frame.

This are some rows from the table.

Symbol	date	stock price
MSFT	2019-02-12	106.8100
MSFT	2019-02-13	106.9000
MSFT	2019-02-14	108.2200
APPL	2018-01-02	168.9873
APPL	2018-01-03	168.9579
APPL	2018-01-04	169.7427

I need to to calculate the ratio {price(i)-price(i-1)} / price(i-1) for the whole table. But I cannot shift my it.

I started with R today, so I'm just struggling a lot.

Magnus Nordmo Magnus Nordmo · Accepted Answer · 2020-12-07T20:09:00

Here is an example using a built-in dataset. I recommend dplyr for these calculations. See the cheatsheet.

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

df <- data(ChickWeight)

df <- as_tibble(ChickWeight)

df <- df %>% 
  group_by(Chick) %>% 
  mutate(change = (weight-first(weight))/first(weight)*100) %>% 
  mutate(change = round(change,2))

^{Created on 2020-12-07 by the reprex package (v0.3.0)}

Computing a Lag on a Table in R

2 Answers

data