Basic example

In this example I will identify all ‘Harami’ patterns, and calculate the Bollinger Bands of the SPDR S&P 500 ETF (SPY).

x <- talib::harami(
  talib::SPY
)

cat(
  "identified patterns:",
  sum(x[, 1] != 0, na.rm = TRUE)
)
#> identified patterns: 35

cat(
  "identified bullish patterns:",
  sum(x[, 1] == 1, na.rm = TRUE)
)
#> identified bullish patterns: 20

cat(
  "identified bearish patterns:",
  sum(x[, 1] == -1, na.rm = TRUE)
)
#> identified bearish patterns: 15

Charting

The Harami pattern can be plotted using talib::chart() with talib::bollinger_bands() to add Bollinger Bands to the chart.

{
  talib::chart(talib::SPY)
  talib::indicator(talib::harami)
  talib::indicator(talib::bollinger_bands)
}

Benchmarks

An often asked question about {talib} in relation to {TTR}, is what it “brings to the table”. Other than Candlestick Patterns and interactive charts, it brings speed and efficiency.

To demonstrate the difference in speed, I will create a univariate price series with 1 million entries.

set.seed(1903)
x <- runif(n = 1e6, min = 100, max = 150)

The univariate series x will be passed into the Bollinger Bands from each package:

bench::mark(
  talib::bollinger_bands(x),
  TTR::BBands(x),
  min_iterations = 10,
  check = FALSE
)[, c(1, 2, 3, 5)]
#> Warning: Some expressions had a GC in every iteration; so filtering is
#> disabled.
#> # A tibble: 2 × 4
#>   expression                     min   median mem_alloc
#>   <bch:expr>                <bch:tm> <bch:tm> <bch:byt>
#> 1 talib::bollinger_bands(x)   6.65ms   9.07ms    22.9MB
#> 2 TTR::BBands(x)             65.12ms  72.42ms   139.3MB

In this benchmark {talib} is faster, and more memory efficient, than {TTR}.

{talib} is still under development, and will most likely not be submitted to CRAN before next year. Until then it can be installed from Github: pak::pak("serkor1/ta-lib-R")

Feel free to stop by the repository here: https://github.com/serkor1/ta-lib-R.

^{Created on 2025-11-16 with reprex v2.1.1}