ggchangepoint

Find where a time series changed with any of 51 methods, and get the same tidy, plottable answer from every one.

🔍 One series, one call, one picture

library(ggchangepoint)
fit <- cpt_detect(Nile, method = "smuce")
fit$changepoints
#> # A tibble: 1 × 5
#>      cp cp_index cp_value ci_lower ci_upper
#>   <int>    <dbl>    <dbl>    <int>    <int>
#> 1    28     1898     1100       26       31
autoplot(fit, show_ci = TRUE)

Five detectors built on different ideas agree:

ggcpt_compare_table(Nile, methods = c("smuce", "not", "wbs", "mosum",
                                      "strucchange"))
#> # A tibble: 5 × 3
#>   method         cp cp_value
#>   <chr>       <int>    <dbl>
#> 1 smuce          28     1100
#> 2 not            28     1100
#> 3 wbs            28     1100
#> 4 mosum          28     1100
#> 5 strucchange    28     1100

🧰 Then ask the next question

You want to know	Call
How big was the change?	cpt_effect(fit)
Where exactly, with no model assumed?	cpt_confint(fit, method = "conformal")
What would have happened without it?	cpt_counterfactual(fit)
Which of my sensors changed?	cpt_attribute(fit)
Did anything change in a year I already had in mind?	cpt_test_at(Nile, when = 1899)
Should I believe this count?	cpt_assumptions(fit)
Which method suits my data?	cpt_recommend(noise = "autocorrelated")
A stretch that returns to normal?	cpt_detect(x, method = "capa")
Watching a live stream?	cpt_monitor("cusum", baseline = b)
What can each of the 51 methods do?	cpt_methods()

🚀 Setup

1. install.packages("ggchangepoint")
2. ggchangepoint::cpt_install_engines(c("core", "regression")) adds the engines this page uses ("all" for every one; three ship with the package).
3. cpt_detect(your_series), then autoplot() the result.

💡 Good to know

For	Do
Data far from unit scale	Use change_in = "meanvar", or standardise first: pelt, binseg and fpop assume unit noise
Autocorrelated series	Check with cpt_assumptions(fit); cpt_robustness(fit) re-runs under another noise model
A date fixed in advance	cpt_test_at() gives an exact test at that date
Gaps in the series	na_action = "omit" keeps the original positions

📚 Learn more

Reference · Choosing a method · Beyond the location · Best practice · Introduction