STATS 326: Applied Time Series
Analysis
STATS 786: Time Series
Forecasting for Data Science
Instructors
2
Shanika Wickramasuriya
Teaches weeks 1–6
Email: s.wickramasuriya@auckland.ac.nz
Office: 303 303
Matt Edwards
Teaches weeks 7–12
Email: matt.edwards@auckland.ac.nz
Office: 303 312
Meeting times
3
Lecture times (NZ time)
Tue 5–6 PM, Owen G Glenn, Room 051
Thu 5–6 PM, Clock Tower - South, Room 039
Tutorial times (NZ time)
Fri 12–1 PM, Science B302, Room G40
Fri 1–2 PM, Science B302, Room G40
At least for the first half of the semester, both lectures and
tutorials will be online.
We will be using tutorial times as office hours. Depending on how
many students turn up, we might increase or decrease the tutorial
hours over the semester.
Learning objectives
1 Use appropriate data visualizations to identify the features present
in time series.
2 Identify the most appropriate time series models for a given
problem.
3 Fit commonly used time series models using R.
4 Interpret and understand the software output for a given time series
model.
5 Perform model selection and cross-validation.
6 Develop computer skills required to forecast time series data.
4
Learning objectives
1 Use appropriate data visualizations to identify the features present
in time series.
2 Identify the most appropriate time series models for a given
problem.
3 Fit commonly used time series models using R.
4 Interpret and understand the software output for a given time series
model.
5 Perform model selection and cross-validation.
6 Develop computer skills required to forecast time series data.
4
Are you interested in learning time series theory...?
Take STATS 726!
References
R. J. Hyndman and G. Athanasopoulos. Forecasting: Principles and
practice. 3rd edition.
R. H. Shumway and D. S. Stoffer. Time series analysis and its
applications: With R examples.
P. J. Brockwell and R. A. Davis. Introduction to time series and
forecasting.
R. J. Hyndman, A. B. Koehler, J. K. Ord and R. D. Snyder. Forecasting
with exponential smoothing: The state space approach.
5
Assessments
Assessment type Due date Percentage
Quizzes See course plan 5% (5× 1%)
Assignments See course plan 15% (3× 5%)
Mid-term test (1 hour) Tuesday 3 May 10%
Group project Friday 3 June 20%
Exam (2 hours) Official exam period 50%
The mid-term test starts at 5 PM (NZ time). Additional time will be given
to upload your solutions on Canvas.
Collaboration is encourage because it opens a space to discuss the
topics that you learn with classmates.
You have to write your own code or/and explanations, and must not
copy another person’s exercise/assignment answers.
6
Assessments
Assessment type Due date Percentage
Quizzes See course plan 5% (5× 1%)
Assignments See course plan 15% (3× 5%)
Mid-term test (1 hour) Tuesday 3 May 10%
Group project Friday 3 June 20%
Exam (2 hours) Official exam period 50%
The mid-term test starts at 5 PM (NZ time). Additional time will be given
to upload your solutions on Canvas.
Collaboration is encourage because it opens a space to discuss the
topics that you learn with classmates.
You have to write your own code or/and explanations, and must not
copy another person’s exercise/assignment answers.
6
Assessments
Assessment type Due date Percentage
Quizzes See course plan 5% (5× 1%)
Assignments See course plan 15% (3× 5%)
Mid-term test (1 hour) Tuesday 3 May 10%
Group project Friday 3 June 20%
Exam (2 hours) Official exam period 50%
The mid-term test starts at 5 PM (NZ time). Additional time will be given
to upload your solutions on Canvas.
Collaboration is encourage because it opens a space to discuss the
topics that you learn with classmates.
You have to write your own code or/and explanations, and must not
copy another person’s exercise/assignment answers. 6
Software and main packages
7
Early preparation
Are you familiar with tidyverse packages?
If you have taken STATS 220 or 369, then you do not have anything
to do.
Otherwise,
I Read the first four chapters of “ModernDive”:
moderndive.netlify.com
8
Main packages
# Data manipulation and plotting functions
library(tidyverse)
# Time series manipulation
library(tsibble)
# Tidy time series data
library(tsibbledata)
# Time series graphics and statistics
library(feasts)
# Forecasting functions
library(fable)
# All of the above
library(fpp3)
9
Main packages
# Data manipulation and plotting functions
library(tidyverse)
# Time series manipulation
library(tsibble)
# Tidy time series data
library(tsibbledata)
# Time series graphics and statistics
library(feasts)
# Forecasting functions
library(fable)
# All of the above
library(fpp3)
9
Installing required packages
install.packages(c(
"tidyverse",
"fpp3"
))
10
Call for class reps
Benefits:
I An important and recognised addition to your resume.
I Improve your leadership skill set.
I Ability to make significant changes to your education.
I At the end of the semester, you will be eligible to receive a class rep
certificate provided you have registered with AUSA.
Responsibilities:
I Elicit feedback from your classmates.
I Attend department student-staff consultative committee meetings
(twice per semester).
I Help to resolve issues that may arise in the class.
11
Department student staff meetings
Undergraduate: STATS 326
Mon 21st Mar (1–2 PM)
Mon 9th May (1–2 PM)
Postgraduate: STATS 786
Thu 24th Mar (12–1 PM)
Fri 29th Apr (12–1 PM)
12