An introduction to time series

Time Series Analysis in Tableau

Christopher Hui

VP of Product, Tracked

What's time series data?

A sequence of observations taken at regular or irregular intervals over time
Granularity references the time period over which the data is collected (e.g. daily, weekly, monthly etc.)

A line chart showing US gasoline prices from 2008 to 2018

Typical time series data might include:

¹ https://www.eia.gov/todayinenergy/detail.php?id=33562

Univariate

Univariate time series analysis refers to the analysis of a single variable over time (e.g. Sales of a product over the past few years)

Multivariate

Multivariate time series analysis is the analysis of multiple variables that are measured over time

A univariate time series line chart showing the price of Copper from 1996 to 2008

A multivariate time series line chart showing the price of multiple commodities (Copper, Coal, Crude Oil) from 1996 to 2008

Continuous time model

Used when data is measured continuously and not at fixed intervals (e.g. stock prices)
These models take into account the exact time interval between measurements

Discrete time model

Used when data is measured and recorded at specific intervals (e.g. daily rainfall)
Does not take into account the time intervals between measurements

An image showing a visualization of a continuous time series

An image showing a visualization of a discrete time series

- DATENAME()
- DATEPART()
- DATETRUNC()

Formatting functions are focused on manipulating the datetime string to suit the granularity of interest. (e.g. year to quarter)

- DATEDIFF() 
- DATEADD()

Exploration functions are focused on calculations using the time series data explicitly (i.e. days between customer purchases)

Time Series Analysis in Tableau