moving_ave: Calculate date-based moving average

In clauswilke/dviz.supp: Supporting materials for Claus Wilke's data visualization book

Description

This function calculates a vector moving averages based on moving date ranges. The result is a vector containing one average per input value, or NA if the full range doesn't exist for a given input value. Not every day within a date range needs to be present.

Usage

moving_ave(date, value, range, center = TRUE)

Arguments

date	Vector of dates
value	Vector of values to average
range	Number of days to average over
center	Boolean, indicates whether the resulting average should be centered in the range (default) or placed at the end of the range

Examples

library(lubridate)

dow_jones_industrial %>% filter(date >= ymd("2009/01/01") & date <= ymd("2010/01/01")) %>%
  mutate(
    close_20d_ave = moving_ave(date, close, 20, center = FALSE),
    close_50d_ave = moving_ave(date, close, 50, center = FALSE),
    close_100d_ave = moving_ave(date, close, 100, center = FALSE)
  ) %>%
  ggplot(aes(date, close)) + 
  geom_line(aes(color = "closing"), size = .25) +
  geom_line(aes(date, close_20d_ave, color = "20d"), size = .75, na.rm = TRUE) +
  geom_line(aes(date, close_50d_ave, color = "50d"), size = .75, na.rm = TRUE) +
  geom_line(aes(date, close_100d_ave, color = "100d"), size = .75, na.rm = TRUE) +
  scale_color_manual(
    values = c(
      closing = "black", 
      `20d` = "#009e73",
      `50d` = "#d55e00",
      `100d` = "#0072b2"
    ),
    breaks = c("closing", "20d", "50d", "100d"),
    labels = c("closing price", "20 day average", "50 day average", "100 day average"),
    name = NULL,
    guide = guide_legend(override.aes = list(size = c(0.25, 0.75, 0.75, 0.75)))
  ) + 
  theme_dviz_grid()

# moving ave vs. LOESS (locally estimated scatterplot smoothing)

dow_jones_industrial %>% filter(date >= ymd("2009/01/01") & date <= ymd("2010/01/01")) %>%
  mutate(
    close_100d_ave = moving_ave(date, close, 100)
  ) %>%
  ggplot(aes(date, close)) + 
  geom_line(aes(color = "closing"), size = .25) +
  geom_line(aes(date, close_100d_ave, color = "100d"), size = .75, na.rm = TRUE) +
  geom_smooth(aes(color = "smooth"), size = .75, na.rm = TRUE, se = FALSE) +
  scale_color_manual(
    values = c(
      closing = "black", 
      `100d` = "#d55e00",
      smooth = "#0072b2"
    ),
    breaks = c("closing", "smooth", "100d"),
    labels = c("closing price", "LOESS smoother", "100 day average"),
    name = NULL,
    guide = guide_legend(override.aes = list(size = c(0.25, 0.75, 0.75)))
  ) + 
  theme_dviz_grid()

clauswilke/dviz.supp documentation built on Aug. 25, 2020, 2:12 a.m.