Market Indicators

Today we are measuring different market indicators using data from the Curated open data on GitHub.

library(tidyverse)
library(praise)

Data

We collect data over time on: gold prices, natural gas prices, oil prices, the yield rate of the US 10y bond, volatility index, and the S & P 500.

gold <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/gold-prices/refs/heads/main/data/monthly-processed.csv'
) |>
  rename(gold_price = Price)

oil <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/oil-prices/refs/heads/main/data/wti-monthly.csv'
) |>
  rename(oil_price = Price)

nat_gas <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/natural-gas/refs/heads/main/data/monthly-processed.csv'
) |>
  rename(gas_price = Price)

vix <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/finance-vix/refs/heads/main/data/vix-monthly.csv'
) |> # volatility index
  rename(vol_idx = Close)

bond <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/bond-yields-us-10y/refs/heads/main/data/monthly.csv'
) |> # bond yields us 10y
  rename(bond_rate = Rate)

temp <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/global-temp/refs/heads/main/data/monthly.csv'
) |>
  mutate(Date = lubridate::as_date(paste0(Year, "-01"))) |>
  rename(temp_diff = Mean) |>
  select(-Year, -Source)
# difference of that month from the average over all of the 20th century

sp <- readr::read_csv(
  'https://raw.githubusercontent.com/datasets/s-and-p-500/refs/heads/main/data/data.csv'
) |>
  select(Date, SP500)

full_data <- gold |>
  full_join(oil) |>
  full_join(nat_gas) |>
  full_join(vix) |>
  full_join(bond) |>
  full_join(temp) |>
  full_join(sp)

Making a single plot

We start by making a single line plot over time, measuring oil prices.

full_data |>
  select(Date, oil_price) |>
  drop_na() |>
  ggplot(aes(x = Date, y = oil_price)) +
  geom_line() +
  theme_bw()

Write a function

The function takes a variable name as the argument and makes a line plot over time for that variable. Note the { } needed due to the non-standard evaluation.

plot_ts <- function(variable, log = TRUE) {
  if (log) {
    full_data |>
      select(Date, {{ variable }}) |>
      drop_na() |>
      ggplot(aes(x = Date, y = {{ variable }})) +
      geom_line() +
      theme_bw() +
      scale_y_log10() +
      labs(x = "date")
  } else {
    full_data |>
      select(Date, {{ variable }}) |>
      drop_na() |>
      ggplot(aes(x = Date, y = {{ variable }})) +
      geom_line() +
      theme_bw() +
      labs(x = "date")
  }
}

plot_ts(SP500, log = TRUE)

plot_ts(SP500, log = FALSE)

Mapping the function

Unforunately, the non-standard evaluation means that the function we wrote cannot be directly mapped. map() will go through the variables vector one at a time (e.g., variables[1], variables[2], etc.), and the name of the variable then needs to be translated into something that can be used in the tidy pipeline.

variables <- c("vol_idx", "gold_price", "oil_price")

map(
  variables,
  \(v, log = TRUE) plot_ts(!!rlang::sym(v), log = log),
  log = FALSE
)

[[1]]

[[2]]

[[3]]

Plotting multiple variables

Instead of looking at a single variable at a time, we might be interested in looking at multiple variables on the same plot. It is a little bit tricky because the variables are meausured on quite different scales, but some are comparable.

full_data |>
  select(Date, bond_rate, temp_diff, SP500) |>
  drop_na() |>
  pivot_longer(
    cols = -Date,
    names_to = "product",
    values_to = "value"
  ) |>
  ggplot(aes(x = Date, y = value, color = product)) +
  geom_line() +
  theme_bw() +
  scale_color_brewer(palette = "Dark2") +
  scale_y_log10() +
  labs(y = "", title = "Comparison over time, log10 scale")

Comparing the US 10y yield bond rate, the S&P 500, and global temperature deviations from the average temperature for all of the 20th century.

Using Shiny

We also played around with Shiny so the user could choose which variable to plot:

praise()

[1] "You are swell!"