time_to_invest.R

# Chris Riederer
# 2019-02-28
# When should you invest your 401k?

library(quantmod)
library(dplyr)
library(ggplot2)

# Pull in S&P data from Yahoo Finance
getSymbols("^GSPC", src="yahoo") 

# Convert to friendly dataframe
df <- as.data.frame(GSPC) %>%
  mutate(date = rownames(.),
         raw_change = GSPC.Close - GSPC.Open,
         change = raw_change / GSPC.Open)

INVEST_AMOUNT = 19000

month_prices <- df %>% 
  group_by(year = year(date), month = month(date)) %>% 
  top_n(-1, wt = date) %>%
  select(date, price = GSPC.Close) %>%
  ungroup()

end_prices <- df %>%
  filter(month(date) == 12) %>%
  group_by(year = year(date)) %>% 
  top_n(1, wt = date) %>%
  select(date, price = GSPC.Close) %>%
  ungroup()  

dff <- data.frame()
for (i in 1:12) {
  tmp_df <- month_prices %>% 
    group_by(year = year(date)) %>% 
    top_n(-i, wt = date) %>%
    ungroup()
  tmp_df <- mutate(tmp_df, strat = i)
  dff <- rbind(dff, tmp_df)
}

avg_price <- dff %>%
  group_by(year, strat) %>%
  summarize(avg_price = mean(price))

foo <- inner_join(avg_price, end_prices, by = "year") %>%
  mutate(growth = price / avg_price - 1)

# Line plots, by year
foo %>%
  ggplot(aes(x=as.factor(strat), y=growth, color=as.factor(year),
             group=year)) +
  geom_point() +
  geom_line() + 
  scale_y_continuous(labels = scales::percent) + 
  scale_x_discrete("Invest in X first months")

# Box plots
foo %>%
  ggplot(aes(x=as.factor(strat), y=growth)) +
  geom_boxplot() +
  scale_y_continuous(labels = scales::percent) + 
  scale_x_discrete("Invest in X first months")

# Line plots, by strat
foo %>%
  ggplot(aes(x=as.factor(year), y=growth, color=as.factor(strat),
             group=strat)) +
  geom_point() +
  # geom_line() + 
  scale_y_continuous(labels = scales::percent) +
  scale_x_discrete("Invest in X first months") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))


# Averages
foo %>%
  group_by(strat) %>%
  summarize(avg_growth = mean(growth)) %>%
  ggplot(aes(x=as.factor(strat), y=avg_growth)) +
  geom_point() + 
  scale_y_continuous(labels = scales::percent) + 
  scale_x_discrete("Invest in X first months")