lab 6 - Time Series Methods

BSMM 8740 Fall 2025

Author

Add your name here

Introduction

In today’s lab, you’ll practice building workflows with recipes, parsnip models, rsample cross validations, and model comparison in the context of time series data.

Getting started

  • To complete the lab, log on to your github account and then go to the class GitHub organization and find the 2025-lab-6-[your github username] repository .

    Create an R project using your 2025-lab-6-[your github username] repository (remember to create a PAT, etc.) and add your answers by editing the 2025-lab-6.qmd file in your repository.

  • When you are done, be sure to: save your document, stage, commit and push your work.

Important

To access Github from the lab, you will need to make sure you are logged in as follows:

  • username: .\daladmin
  • password: Business507!

Remember to (create a PAT and set your git credentials)

  • create your PAT using usethis::create_github_token() ,
  • store your PAT with gitcreds::gitcreds_set() ,
  • set your username and email with
    • usethis::use_git_config( user.name = ___, user.email = ___)
Note

The code below will install (if necessary) and load the packages needed in today’s exercises

Packages

Package installs

If you install packages outside of the ones below, remove your install.packages(.) code before submitting your solutions.

Leaving instructions that will install packages may cause your code to fail, but more importantly, it can corrupt the TA/instructor’s system.

The Data

Today we will be using electricity demand data, based on a paper by James W Taylor:

Taylor, J.W. (2003) Short-term electricity demand forecasting using double seasonal exponential smoothing. Journal of the Operational Research Society, 54, 799-805.

The data can be found in the timetk package as timetk::taylor_30_min, a tibble with dimensions: 4,032 x 2

  • date: A date-time variable in 30-minute increments

  • value: Electricity demand in Megawatts

data <- timetk::taylor_30_min

Exercise 1: EDA

Plot the data using the functions timetk::plot_time_series, timetk::plot_acf_diagnostics (using 100 lags), and timetk::plot_seasonal_diagnostics.

YOUR ANSWER: 
# timetk::plot_time_series
# timetk::plot_acf_diagnostics
# timetk::plot_seasonal_diagnostics

Exercise 2: Time scaling

The raw data has 30 minute intervals between data points. Downscale the data to 60 minute intervals, using timetk::summarise_by_time, revising the electricity demand (value) variable by adding the two 30-minute intervals in each 60-minute interval. Assign the downscaled data to the variable taylor_60_min.

YOUR ANSWER: 
# downscale the data (down to a lower frequency of measurement)
set.seed(8740)
taylor_60_min <- ?

Exercise 3: Training and test datasets

  1. Split the new (60 min) time series into training and test sets using timetk::time_series_split
    1. set the training period (‘initial’) to ‘2 months’ and the assessment period to ‘1 weeks’
  2. Prepare the data resample specification with timetk::tk_time_series_cv_plan() and plot it with timetk::plot_time_series_cv_plan
  3. Separate the training and test data sets using rsample .
YOUR ANSWER: 
# split
# plot splits
# separate into train and test data sets

Exercise 4: recipes

  1. Create a base recipe (base_rec) using the formula value ~ date and the training data. This will be used for non-regression models

  2. Create a recipe (lm_rec) using the formula value ~ . and the training data. This will be used for regression models. For this recipe:

    1. add time series signature features using timetk::step_timeseries_signature with the appropriate argument,
    2. add a step to select the columns value, date_index.num, date_month.lbl, date_wday.lbl, date_hour ,
    3. add a normalization step targeting date_index.num ,
    4. add a step to mutate date_hour, changing it to a factor,
    5. add a step to one-hot encode nominal predictors.
YOUR ANSWER: 
base_rec <- ?
  
lm_rec <- ?

Exercise 5 models

Now we will create a several models to estimate electricity demand, as follows

  1. Create a model specification for an exponential smoothing model using engine ‘ets’
  2. Create a model specification for an arima model using engine ‘auto_arima’
  3. Create a model specification for a linear model using engine ‘glmnet’ and penalty = 0.02, mixture = 0.5
YOUR ANSWER: 
model_ets <- ?
  
model_arima <- ?
  
model_lm <- ?

Exercise 6 model fitting

Create a workflow for each model using workflows::workflow.

  1. Add a recipe to the workflow
    1. the linear model uses the lm_rec recipe created above
    2. the ets and arima models use the base_rec recipe created above
  2. Add a model to each workflow
  3. Fit with the training data
YOUR ANSWER: 
workflow_fit_ets <- ?
  
workflow_fit_arima <- ?
  
workflow_fit_lm <- ?

This is a good place to save, stage, commit, and push changes to your remote lab repo on GitHub. Click the checkbox next to each file in the Git pane to stage the updates you’ve made, write an informative commit message, and push. After you push the changes, the Git pane in RStudio should be empty.

Exercise 7: calibrate

In this exercise we’ll use the testing data with our fitted models.

  1. Create a table with the fitted workflows using modeltime::modeltime_table
  2. Using the table you just created, run a calibration on the test data with the function modeltime::modeltime_calibrate.
  3. Compare the accuracy of the models using the modeltime::modeltime_accuracy() on the results of the calibration
YOUR ANSWER: 
model_tbl <- modeltime::modeltime_table( ? )

calibration_tbl <- ?
  
calibration_tbl |> ?
Important

Which is the best model by the rmse metric?

Exercise 8: forecast - training data

Use the calibration table with modeltime::modeltime_forecast to graphically compare the fits to the training data with the observed values.

YOUR ANSWER:

Exercise 9: forecast - future

Now refit the models using the full data set (using the calibration table and modeltime::modeltime_refit). Save the result in the variable refit_tbl.

  1. Use the refit data in the variable refit_tbl, along with modeltime::modeltime_forecast and argument h = ‘2 weeks’ (remember to also set the actual_data argument). This will use the models to forecast electricity demand two weeks into the future.
  2. Plot the forecast with modeltime::plot_modeltime_forecast
YOUR ANSWER: 
refit_tbl <-  ?

Exercise 10: Seasonal Time Series Modeling for Retail Forecasting

You are working as a Senior Business Analyst for “TechMart,” a consumer electronics retailer with 50 stores across North America. The company is planning its inventory strategy for 2025 and needs accurate forecasts of monthly sales. The CFO has specifically requested that you use advanced decomposition methods that can handle the complex seasonal patterns in electronics retail.

You have been provided with 60 months of historical sales data (Jan 2019 - Dec 2023) showing strong monthly seasonality (holiday shopping peaks) and an underlying growth trend due to market expansion.

# retail data
retail_dat <- readr::read_csv('data/TechMart_sales.csv',show_col_types = FALSE)
# Display data summary
cat("TechMart Monthly Sales Data Summary:\n")
TechMart Monthly Sales Data Summary:
cat("====================================\n")
====================================
cat("Period:", format(min(retail_dat$date), "%b %Y"), "to", 
    format(max(retail_dat$date), "%b %Y"), "\n")
Period: Jan 2019 to Dec 2023 
cat("Observations:", nrow(retail_dat), "\n")
Observations: 60 
print(summary(retail_dat$monthly_sales))
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  2.979   4.364   5.716   5.843   6.830  10.585

Using the modeltime and parsnip frameworks,

  1. build and evaluate a Seasonal and Trend decomposition using Loess with ETS (STLM-ETS) model.

  2. compare your model against the SNAIVE model (baseline) and an ETS(A,A,A) model.

  3. pick the best model based on rmse and show the corresponding mape. Also express the rmse as a percent improvement over the baseline.

  4. forecast sales one month ahead, and extract the forecast for December.

YOUR ANSWER: 
# SHOW ALL YOUR WORK

You’re done and ready to submit your work! Save, stage, commit, and push all remaining changes. You can use the commit message “Done with Lab 6!” , and make sure you have committed and pushed all changed files to GitHub (your Git pane in RStudio should be empty) and that all documents are updated in your repo on GitHub.

Submission

I will pull (copy) everyone’s repository submissions at 5:00pm on the Sunday following class, and I will work only with these copies, so anything submitted after 5:00pm will not be graded. (don’t forget to commit and then push your work!)

Grading

Total points available: 30 points.