Pipe-friendly function that adds trend columns to a tibble or data.frame. Designed for exploratory analysis of monthly and quarterly economic time series. Supports multiple trend extraction methods and handles grouped data.
Usage
augment_trends(
data,
date_col = "date",
value_col = "value",
group_cols = NULL,
group_vars = NULL,
methods = "stl",
frequency = NULL,
suffix = NULL,
window = NULL,
smoothing = NULL,
band = NULL,
align = NULL,
params = list(),
.quiet = FALSE
)Arguments
- data
A
data.frame,tibble,data.table, ortsibblecontaining the time series data. Tsibble support requires the optional tsibble package.- date_col
Name of the date column. Defaults to
"date". Must be of classDatefor data frames. For tsibbles, defaults to the index and must name that index when supplied.- value_col
Name of the value column(s). Defaults to
"value". Must benumeric. A character vector of length > 1 is accepted; trends are extracted for each column and namedtrend_{method}_{col}(e.g.trend_stl_consumption).- group_cols
Optional grouping variables for multiple time series. Can be a character vector of column names. For tsibbles, defaults to the key and must match it when supplied.
- group_vars
Deprecated. Use
group_colsinstead.- methods
Character vector of trend methods. Options:
"hp","bk","cf","ma","stl","loess","spline","poly","bn","ucm","hamilton","spencer","henderson","ewma","wma","triangular","kernel","kalman","median","gaussian". Default is"stl".- frequency
The frequency of the series. Supports values from 1 (annual) to 365 (daily). Auto-detected for data frames; a tsibble's
yearmonthoryearquarterindex supplies 12 or 4. ADateindex uses the usual detection. Other tsibble index classes are not supported.- suffix
Optional suffix for trend column names. If NULL, uses method names.
- window
Unified window/period parameter for moving average methods (ma, wma, triangular, stl, ewma, median, gaussian, henderson). Must be positive. If NULL, uses frequency-appropriate defaults. For EWMA, the window is converted to the smoothing factor via
alpha = 2 / (window + 1). Cannot be used simultaneously withsmoothingfor EWMA method. Forma,median, andhendersonmethods, a numeric vector is accepted (e.g.,c(9, 13, 23)), which adds one column per window value namedtrend_henderson_9,trend_henderson_13, etc. Other methods ignore extra values (with a warning).- smoothing
Unified smoothing parameter for smoothing methods (hp, loess, spline, ewma, kernel, kalman). For hp: use large values (1600+) or small values (0-1) that get converted. For EWMA: specifies the alpha parameter (0-1) for traditional exponential smoothing. Cannot be used simultaneously with
windowfor EWMA method. For kernel: multiplier of optimal bandwidth (1.0 = optimal, <1 = less smooth, >1 = more smooth). For kalman: a finite, positive ratio of measurement to process noise (higher = more smoothing). An explicit noise variance inparamsdetermines the other variance from this ratio. If both variances are supplied, they take precedence oversmoothing. Without a ratio, unspecified measurement and process variances default to 0.1 and 0.01 times the series variance. For others: typically 0-1 range.- band
Unified band parameter for bandpass filters (bk, cf). Provide as
c(low, high): the shortest and longest cycle to remove, in periods of the series (months for monthly data). Both values must be positive. Defaults to cycles of 1.5 to 8 years:c(6, 32)for quarterly data,c(18, 96)for monthly, andc(2, 8)for annual.- align
Unified alignment parameter for moving average methods (ma, wma, triangular, gaussian). Valid values:
"center"(default, uses surrounding values),"right"(causal, uses past values only),"left"(anti-causal, uses future values only). Note: triangular only supports"center"and"right". If NULL, uses"center"as default.- params
Optional list of method-specific parameters for fine control.
- .quiet
If
TRUE, suppress informational messages.
Value
A tibble with original data plus trend columns named trend_{method} or
trend_{method}_{suffix} if suffix is provided. Rows come back in the
order they were supplied in. A tsibble input returns a tsibble with its
index class and key preserved.
Details
This function is designed for monthly (frequency = 12) and quarterly (frequency = 4) economic data, and the defaults for each method follow the conventions for those frequencies.
For grouped data, the function applies trend extraction to each group separately,
maintaining the original data structure while adding trend columns.
For tsibbles, only Date, yearmonth, and yearquarter indices are
supported; the existing missing-period rules apply after conversion to
calendar dates.
Examples
# Simple STL decomposition on quarterly GDP construction data
gdp_construction |> augment_trends(value_col = "index")
#> Auto-detected quarterly (4 obs/year)
#> # A tibble: 124 × 3
#> date index trend_stl
#> <date> <dbl> <dbl>
#> 1 1995-01-01 100 102.
#> 2 1995-04-01 100 101.
#> 3 1995-07-01 100 100.
#> 4 1995-10-01 100 99.4
#> 5 1996-01-01 97.8 101.
#> 6 1996-04-01 101. 102.
#> 7 1996-07-01 107. 103.
#> 8 1996-10-01 103. 104.
#> 9 1997-01-01 101. 106.
#> 10 1997-04-01 108. 109.
#> # ℹ 114 more rows
# Multiple smoothing methods with unified parameter
gdp_construction |>
augment_trends(
value_col = "index",
methods = c("hp", "loess", "ewma"),
smoothing = 0.3
)
#> Auto-detected quarterly (4 obs/year)
#> # A tibble: 124 × 5
#> date index trend_hp trend_loess trend_ewma
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 1995-01-01 100 99.1 97.7 100
#> 2 1995-04-01 100 100. 99.1 100
#> 3 1995-07-01 100 101. 100. 100
#> 4 1995-10-01 100 102. 102. 100
#> 5 1996-01-01 97.8 103. 103. 99.3
#> 6 1996-04-01 101. 104. 104. 99.8
#> 7 1996-07-01 107. 104. 105. 102.
#> 8 1996-10-01 103. 105. 105. 102.
#> 9 1997-01-01 101. 106. 106. 102.
#> 10 1997-04-01 108. 107. 107. 104.
#> # ℹ 114 more rows
# Moving averages with unified window on monthly data
vehicles |>
tail(60) |>
augment_trends(
value_col = "production",
methods = c("ma", "wma", "triangular"),
window = 8
)
#> Auto-detected monthly (12 obs/year)
#> # A tibble: 60 × 5
#> date production trend_ma trend_wma trend_triangular
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 2021-01-01 180904 NA NA NA
#> 2 2021-02-01 186718 NA NA NA
#> 3 2021-03-01 208801 NA NA NA
#> 4 2021-04-01 191853 NA 188418. 193806.
#> 5 2021-05-01 206221 188457. 181049. 190493.
#> 6 2021-06-01 191571 185917 177322. 184845.
#> 7 2021-07-01 174739 182853. 175420. 179033.
#> 8 2021-08-01 178900 182548. 184002. 176951.
#> 9 2021-09-01 156803 179890. 173298. 174482.
#> 10 2021-10-01 170178 173445. 169534. 173903.
#> # ℹ 50 more rows
# Economic indicators with different methods
ibcbr |>
tail(48) |>
augment_trends(
value_col = "index",
methods = c("median", "kalman", "kernel"),
window = 9,
smoothing = 0.15
)
#> Auto-detected monthly (12 obs/year)
#> # A tibble: 48 × 5
#> date index trend_median trend_kalman trend_kernel
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 2022-01-01 91.8 91.8 92.4 91.8
#> 2 2022-02-01 95.6 95.6 96.1 95.6
#> 3 2022-03-01 105. 100. 103. 105.
#> 4 2022-04-01 100. 100. 100. 100.
#> 5 2022-05-01 99.7 100. 99.8 99.7
#> 6 2022-06-01 99.3 100. 99.8 99.3
#> 7 2022-07-01 104. 100. 104. 104.
#> 8 2022-08-01 105. 100. 104. 104.
#> 9 2022-09-01 101. 100.0 101. 101.
#> 10 2022-10-01 100. 100.0 100. 100.
#> # ℹ 38 more rows
# Moving average with right alignment (causal filter)
vehicles |>
tail(60) |>
augment_trends(
value_col = "production",
methods = "ma",
window = 12,
align = "right"
)
#> Auto-detected monthly (12 obs/year)
#> # A tibble: 60 × 3
#> date production trend_ma
#> <date> <dbl> <dbl>
#> 1 2021-01-01 180904 NA
#> 2 2021-02-01 186718 NA
#> 3 2021-03-01 208801 NA
#> 4 2021-04-01 191853 NA
#> 5 2021-05-01 206221 NA
#> 6 2021-06-01 191571 NA
#> 7 2021-07-01 174739 NA
#> 8 2021-08-01 178900 NA
#> 9 2021-09-01 156803 NA
#> 10 2021-10-01 170178 NA
#> # ℹ 50 more rows
# Advanced: fine-tune specific methods
electric |>
tail(72) |>
augment_trends(
value_col = "consumption",
methods = "median",
window = 7
)
#> Auto-detected monthly (12 obs/year)
#> # A tibble: 72 × 3
#> date consumption trend_median
#> <date> <dbl> <dbl>
#> 1 2020-01-01 12909 12530
#> 2 2020-02-01 12383 12432
#> 3 2020-03-01 12432 12383
#> 4 2020-04-01 12318 12318
#> 5 2020-05-01 11756 11852
#> 6 2020-06-01 11396 11852
#> 7 2020-07-01 11707 11852
#> 8 2020-08-01 11852 11852
#> 9 2020-09-01 12240 12240
#> 10 2020-10-01 13090 12778
#> # ℹ 62 more rows
# Multiple MA windows in a single call (adds trend_ma_3, trend_ma_6, trend_ma_12)
vehicles |>
tail(60) |>
augment_trends(
value_col = "production",
methods = "ma",
window = c(3, 6, 12)
)
#> Auto-detected monthly (12 obs/year)
#> # A tibble: 60 × 5
#> date production trend_ma_3 trend_ma_6 trend_ma_12
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 2021-01-01 180904 NA NA NA
#> 2 2021-02-01 186718 192141 NA NA
#> 3 2021-03-01 208801 195791. NA NA
#> 4 2021-04-01 191853 202292. 193831. NA
#> 5 2021-05-01 206221 196548. 192666. NA
#> 6 2021-06-01 191571 190844. 187681 NA
#> 7 2021-07-01 174739 181737. 181542. 185005.
#> 8 2021-08-01 178900 170147. 177244. 181965.
#> 9 2021-09-01 156803 168627 177075. 179091.
#> 10 2021-10-01 170178 167768. 176178. 176611.
#> # ℹ 50 more rows
# Preserve a tsibble's index and key (if tsibble is installed)
if (requireNamespace("tsibble", quietly = TRUE)) {
quarterly <- gdp_construction
quarterly$date <- tsibble::yearquarter(quarterly$date)
quarterly <- tsibble::as_tsibble(quarterly, index = date)
augment_trends(quarterly, value_col = "index", methods = "hp")
}
#> # A tsibble: 124 x 3 [1Q]
#> date index trend_hp
#> <qtr> <dbl> <dbl>
#> 1 1995 Q1 100 101.
#> 2 1995 Q2 100 101.
#> 3 1995 Q3 100 102.
#> 4 1995 Q4 100 103.
#> 5 1996 Q1 97.8 103.
#> 6 1996 Q2 101. 104.
#> 7 1996 Q3 107. 104.
#> 8 1996 Q4 103. 105.
#> 9 1997 Q1 101. 106.
#> 10 1997 Q2 108. 106.
#> # ℹ 114 more rows