Defining T0 in my program - r

Question

Here's a small program I'm making, to eventually get a final graph. I have 2 separate data sets. One is called T0 and the second one contains all the data I have. I want this program to get the T0 values from the the first data frame, then it searches about the maximum price in the 3 years before and the 3 years after the T0 year.
In essence, my program is going to assign T0 values that I chose arbitrarily. Then it will search automatically in my database for the maximum price in each year except the t0 year.
The problem I'm facing, is with the implementation of T0 values in the schedule. It just does not come out right when I run my code.
The problem apparently has to do with the way I'm defining T0. Should I use a for loop? or is there a small tweak I'm missing?
Final result wanted:
Data Base Example:
T0data:
structure(list(Company = structure(1:3, .Label = c("Amazon",
"Cisco", "McDonald's"), class = "factor"), Year = c(2011L, 2008L,
2013L), Price = c(182, 21.82, 95.15)), .Names = c("Company",
"Year", "Price"), row.names = c(NA, 3L), class = "data.frame")
All Data:
structure(list(Company = structure(c(1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 3L, 3L, 3L, 3L,
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L
), .Label = c("Amazon", "Cisco", "McDonald's"), class = "factor"),
Year = c(2008L, 2008L, 2008L, 2008L, 2009L, 2009L, 2010L,
2010L, 2010L, 2011L, 2011L, 2012L, 2012L, 2013L, 2013L, 2014L,
2014L, 2014L, 2008L, 2010L, 2010L, 2010L, 2011L, 2011L, 2012L,
2012L, 2013L, 2013L, 2014L, 2014L, 2014L, 2015L, 2015L, 2016L,
2016L, 2016L, 2005L, 2005L, 2005L, 2006L, 2006L, 2007L, 2007L,
2007L, 2008L, 2008L, 2009L, 2009L, 2009L, 2010L, 2010L, 2011L,
2011L, 2011L), Price = c(91L, 77L, 81L, 87L, 63L, 88L, 110L,
75L, 117L, 170L, 190L, 215L, 245L, 316L, 275L, 330L, 378L,
390L, 55L, 62L, 66L, 65L, 72L, 98L, 93L, 88L, 99L, 101L,
94L, 103L, 96L, 99L, 116L, 112L, 123L, 113L, 19L, 17L, 18L,
20L, 19L, 26L, 31L, 27L, 24L, 21L, 14L, 22L, 18L, 26L, 22L,
14L, 16L, 15L)), .Names = c("Company", "Year", "Price"), class = "data.frame", row.names = c(NA,
-54L))
My code:
library(data.table)
T0data<- read.csv(file = "C:/Users/My first file.csv", header = TRUE )
Alldata<- read.csv(file = "C:/Users/My second file.csv", header = TRUE )
d<-Alldata
setDT(d)
year_zero <- T0data$Year
# Filter to include year_zero +/- 3 years and get Best result per company per year
d <- d[Year >= year_zero - 3 & Yeae <= year_zero + 3,
.(Best_Result = max(Price, na.rm = TRUE)), by = .(Company, Year)]
# Add T as interval to year_zero (and convert to factor in order to get all
# values from 3 to 3
d[, "T" := factor(Year - year_zero, levels = seq(-3, 3), ordered = TRUE)]
# Cast to wide format (fill missing values with NA)
dcast(d, Company ~T, value.var = "Best_Result", drop = FALSE)
# Cast to wide format (fill missing values with "")
dcast(d, Company~T, value.var = "Best_Result", drop = FALSE, fun.aggregate = paste0,
fill = "")

Answer 1

Here's a solution that uses dplyr / tidyr packages from the tidyverse, rather than data.table, but it should do the job:
library(dplyr); library(tidyr)
T0.modified <- T0data %>%
# create year range based on each company's T0 year
mutate(Year.M1 = Year - 1,
Year.M2 = Year - 2,
Year.M3 = Year - 3,
Year.P1 = Year + 1,
Year.P2 = Year + 2,
Year.P3 = Year + 3) %>%
# convert to long format, match with Alldata based on both company & year
gather(reference.year, actual.year, -Company, -Price) %>%
left_join(Alldata, by = c("Company" = "Company", "actual.year" = "Year")) %>%
# keep T0 price for year T0, & use matched prices for all other years
mutate(Price = ifelse(reference.year == "Year", Price.x, Price.y)) %>%
# take maximum of all matched prices for each company each year
group_by(Company, reference.year) %>%
summarise(Price = max(Price)) %>%
ungroup() %>%
# order reference.year for correct sequence in ggplot's x-axis
mutate(reference.year = factor(reference.year,
levels = c("Year.M3", "Year.M2", "Year.M1", "Year",
"Year.P1", "Year.P2", "Year.P3"),
labels = c("T-3", "T-2", "T-1", "T0", "T+1", "T+2", "T+3")))
Resulting plot:
library(ggplot2)
ggplot(T0.modified,
aes(x = reference.year, y = Price, group = Company, color = Company)) +
geom_line(aes()) +
xlab("Year") + theme_bw()
Edit adding average for each year using stat_summary:
ggplot(T0.modified,
aes(x = reference.year, y = Price, group = Company, color = Company)) +
geom_line(aes()) +
xlab("Year") + theme_bw() +
stat_summary(fun.y = mean, geom = "line", group = 1,
linetype = 2, size = 1.5, colour = "grey") +
annotate("label", x = 7, y = 200, label = "Average",
fill = "grey", alpha = 0.5, hjust = 1)