Package index • timetk

Plotting Time Series

Detect relationships through visualizations

Time Series Plotting

plot_time_series(): Interactive Plotting for One or More Time Series

plot_time_series_boxplot(): Interactive Time Series Box Plots

plot_time_series_regression(): Visualize a Time Series Linear Regression Formula

Correlation, Seasonalilty, & Anomaly Plotting

plot_acf_diagnostics(): Visualize the ACF, PACF, and CCFs for One or More Time Series

plot_anomaly_diagnostics(): Visualize Anomalies for One or More Time Series

plot_seasonal_diagnostics(): Visualize Multiple Seasonality Features for One or More Time Series

plot_stl_diagnostics(): Visualize STL Decomposition Features for One or More Time Series

Time Series Data Wrangling Operations

Extension for dplyr for time-series data manipulations

Data Frame Operations

summarise_by_time() summarize_by_time(): Summarise (for Time Series Data)

mutate_by_time(): Mutate (for Time Series Data)

pad_by_time(): Insert time series rows with regularly spaced timestamps

filter_by_time(): Filter (for Time-Series Data)

filter_period(): Apply filtering expressions inside periods (windows)

slice_period(): Apply slice inside periods (windows)

condense_period(): Convert the Period to a Lower Periodicity (e.g. Go from Daily to Monthly)

future_frame(): Make future time series from existing

Anomaly Detection

anomalize(): Automatic group-wise Anomaly Detection

plot_anomalies() plot_anomalies_decomp() plot_anomalies_cleaned(): Visualize Anomalies for One or More Time Series

Function Operations

slidify(): Create a rolling (sliding) version of any function

Vector Operations

between_time(): Between (For Time Series): Range detection for date or date-time sequences

add_time() subtract_time() `%+time%` `%-time%`: Add / Subtract (For Time Series)

Time Series Features

Tidy integration with tsfeatures

tk_tsfeatures(): Time series feature matrix (Tidy)

Augment Operations (Quickly Add Many Features)

Add multiple columns to the original data. Respects dplyr groups.

tk_augment_timeseries_signature(): Add many time series features to the data

tk_augment_holiday_signature(): Add many holiday features to the data

tk_augment_slidify(): Add many rolling window calculations to the data

tk_augment_differences(): Add many differenced columns to the data

tk_augment_lags() tk_augment_leads(): Add many lags to the data

tk_augment_fourier(): Add many fourier series to the data

Vectorized Transformations

Use with mutate to apply vectorized transformations to time series data

box_cox_vec() box_cox_inv_vec() auto_lambda(): Box Cox Transformation

diff_vec() diff_inv_vec(): Differencing Transformation

lag_vec() lead_vec(): Lag Transformation

standardize_vec() standardize_inv_vec(): Standardize to Mean 0, Standard Deviation 1 (Center & Scale)

normalize_vec() normalize_inv_vec(): Normalize to Range (0, 1)

fourier_vec(): Fourier Series

log_interval_vec() log_interval_inv_vec(): Log-Interval Transformation for Constrained Interval Forecasting

slidify_vec(): Rolling Window Transformation

smooth_vec(): Smoothing Transformation using Loess

ts_clean_vec(): Replace Outliers & Missing Values in a Time Series

ts_impute_vec(): Missing Value Imputation for Time Series

Feature Engineering Operations (Recipe Steps)

Preprocessing & feature engineering operations for use with recipes and the tidymodels ecosystem

Engineered Features

step_timeseries_signature() tidy(<step_timeseries_signature>): Time Series Feature (Signature) Generator

step_holiday_signature() tidy(<step_holiday_signature>): Holiday Feature (Signature) Generator

step_fourier() tidy(<step_fourier>): Fourier Features for Modeling Seasonality

Lags & Diffs

See recipes::step_lag() for lagged features.

step_diff() tidy(<step_diff>): Create a differenced predictor

Smoothing & Rolling

step_smooth() tidy(<step_smooth>): Smoothing Transformation using Loess

step_slidify() tidy(<step_slidify>): Slidify Rolling Window Transformation

step_slidify_augment() tidy(<step_slidify_augment>): Slidify Rolling Window Transformation (Augmented Version)

Variance Reduction

See recipes::step_log() for log transformation.

step_box_cox() tidy(<step_box_cox>): Box-Cox Transformation using Forecast Methods

step_log_interval() tidy(<step_log_interval>): Log Interval Transformation for Constrained Interval Forecasting

Add Rows to a Time series

step_ts_pad() tidy(<step_ts_pad>): Pad: Add rows to fill gaps and go from low to high frequency

Imputation & Outlier Cleaning

See recipes::step_rollimpute() for rolling imputation.

step_ts_impute() tidy(<step_ts_impute>): Missing Data Imputation for Time Series

step_ts_clean() tidy(<step_ts_clean>): Clean Outliers and Missing Data for Time Series

Cross Validation Operations (Rsample & Tune)

Resampling for time series cross validation using rsamples

Time Series Cross Validation (Resample Sets)

time_series_split(): Simple Training/Test Set Splitting for Time Series

time_series_cv(): Time Series Cross Validation

Cross Validation Plan Visualization (Resample Sets)

Uses the output of time_series_cv or rsample::rolling_origin

plot_time_series_cv_plan(): Visualize a Time Series Resample Plan

tk_time_series_cv_plan(): Time Series Resample Plan Data Preparation

Index Operations

Extract and check the date or date-time index.

tk_index() has_timetk_idx(): Extract an index of date or datetime from time series objects, models, forecasts

Make Operations

Make time series sequences.

tk_make_timeseries(): Intelligent date and date-time sequence creation

tk_make_future_timeseries(): Make future time series from existing

tk_make_holiday_sequence() tk_make_weekend_sequence() tk_make_weekday_sequence(): Make daily Holiday and Weekend date sequences

Get Operations

Get summaries, frequency, and signatures from the time series index.

tk_get_timeseries_signature() tk_get_timeseries_summary(): Get date features from a time-series index

tk_get_holiday_signature() tk_get_holidays_by_year(): Get holiday features from a time-series index

tk_get_frequency() tk_get_trend(): Automatic frequency and trend calculation from a time series index

tk_get_timeseries_unit_frequency(): Get the timeseries unit frequency for the primary time scales

tk_get_timeseries_variables(): Get date or datetime variables (column names)

Diagnostic Operations

These power the time series plotting functions

tk_summary_diagnostics(): Group-wise Time Series Summary

tk_anomaly_diagnostics(): Automatic group-wise Anomaly Detection by STL Decomposition

tk_acf_diagnostics(): Group-wise ACF, PACF, and CCF Data Preparation

tk_seasonal_diagnostics(): Group-wise Seasonality Data Preparation

tk_stl_diagnostics(): Group-wise STL Decomposition (Season, Trend, Remainder)

Conversion Operations

Functions for converting between common time series formats.

tk_tbl(): Coerce time-series objects to tibble.

tk_ts() tk_ts_(): Coerce time series objects and tibbles with date/date-time columns to ts.

tk_xts() tk_xts_(): Coerce time series objects and tibbles with date/date-time columns to xts.

tk_zoo() tk_zoo_(): Coerce time series objects and tibbles with date/date-time columns to xts.

tk_zooreg() tk_zooreg_(): Coerce time series objects and tibbles with date/date-time columns to ts.

Time Scale Template

The timescale template is used to automate frequency and trendcycle calculations.

set_tk_time_scale_template() get_tk_time_scale_template() tk_time_scale_template(): Get and modify the Time Scale Template

Time Series Datasets

Time series from various forecasting competitions. Domains include economic, retail, and web (google analytics)

m4_hourly: Sample of 4 Hourly Time Series Datasets from the M4 Competition

m4_daily: Sample of 4 Daily Time Series Datasets from the M4 Competition

m4_weekly: Sample of 4 Weekly Time Series Datasets from the M4 Competition

m4_monthly: Sample of 4 Monthly Time Series Datasets from the M4 Competition

m4_quarterly: Sample of 4 Quarterly Time Series Datasets from the M4 Competition

m4_yearly: Sample of 4 Yearly Time Series Datasets from the M4 Competition

walmart_sales_weekly: Sample Time Series Retail Data from the Walmart Recruiting Store Sales Forecasting Competition

wikipedia_traffic_daily: Sample Daily Time Series Data from the Web Traffic Forecasting (Wikipedia) Competition

bike_sharing_daily: Daily Bike Sharing Data

taylor_30_min: Half-hourly electricity demand

FANG: Stock prices for the "FANG" stocks.

Date Utilities

parse_date2() parse_datetime2(): Fast, flexible date and datetime parsing

is_date_class(): Check if an object is a date class

Package Information

timetk timetk-package: timetk: Time Series Analysis in the Tidyverse