LoginSign Up
← Back to all sparks
T

trendseries

ANALYTICS
Velocity3.8

R package trendseries by viniciusoike — release notes from GitHub.

github.com/viniciusoike/trendseries ↗

A trend-extraction toolkit grows a full decomposition engine, seasonal components and all.

time-serieseconometricsr-packageseasonal-decompositiontrend-extractionreproducibility
◆Current state
trendseries extracts trends from economic time series through two pipe-friendly functions, backed by an unusually broad method set — Hodrick-Prescott in one- and two-sided variants, Baxter-King, Christiano-Fitzgerald, Hamilton regression, Beveridge-Nelson, unobserved components, plus the moving average and smoothing family. The 1.4 release adds decomposition proper: an exported decompose_series() that splits a series into trend, seasonal, and remainder across five methods and guarantees the components add back to the original values.
◆Where it's heading
The package is moving from breadth of methods to rigour about what those methods produce. Recent work has been about defaults and guarantees rather than new filters: the unobserved components model now derives its signal-to-noise ratios from Hodrick-Prescott lambdas so the default output is economically interpretable, decomposition carries an exact additive identity, and a log transform gives a uniform multiplicative variant across every method. Naming is being tidied in the same spirit, with group_vars deprecated in favour of group_cols. Side-by-side method comparison — passing several methods and getting each one's components as separate columns — suggests an audience that treats method choice as a research question rather than a setting.
◆Prediction
Expect the comparison and diagnostic side to keep developing, since the package now produces multiple decompositions of the same series and offers no ranking between them; the entries give no indication of new filters being queued.

◆Recent moves

  1. 15d ago

    Decomposition becomes a first-class operation, five methods deep

    ⚡ SPARK

    The package's output type changes: alongside trend columns it now returns seasonal and remainder components, via an exported decompose_series() covering STL, regression, classical moving averages, a state-space basic structural model, and X-13ARIMA-SEATS. It also folds in the 1.3.0 development series that never reached CRAN, so this release carries more than its version gap suggests.

    View source ↗
  2. 3mo ago

    Multi-column trends and economically grounded UCM defaults

    augment_trends() accepts several value columns at once, naming outputs by method and column, which matters for anyone running the same filter across a panel. The more substantive change is under the hood: the unobserved components model now uses fixed variance components with signal-to-noise ratios derived from Hodrick-Prescott lambdas, so its default trends are smoother and interpretable without tuning. group_vars is deprecated in favour of group_cols.

    View source ↗
  3. 10mo ago

    First production release with 21 trend extraction methods

    The baseline the later releases build on: 21 methods spanning econometric filters, moving averages, smoothing and signal processing, behind a deliberately small two-function API. The breadth established here is what makes the later work — better defaults, exact identities, side-by-side comparison — the useful next problem.

    View source ↗