Data Science: Where Are Canadians Travelling?

Posted by

Where Are Canadians Travelling To?

Despite living in the second largest country in the world, sometimes you just need to get away from things, eh? Statistics Canada provides a yearly summary table depicting the top travel destinations for Canadians in the form of the Travel Survey of Residents of Canada (TSRC). The visualizations give us some insight into our habits and how global events affect our travel plans. Visualization have all been written in R, largely using the gpplot2 package.

All code and a cleaned up data file can be found on my GitHub.

A Short Primer on the Data

I have compiled all the summary tables into a single file, as well as a CPI table and a Canadian Population table. The data is presented in yearly intervals, thus line charts will contain a single point for each year. The data collected spans from 2001-2015, and represents the top 15 countries visited by residents of Canada.

The travel data has five data entries:
– Year
– Country
– Number of Visits
– Nights Spent
– Dollar Spent in Canadian Dollars

For those with questions about the survey methodology, data imputation strategy and the assumptions (i.e. visiting multiple countries), please check out the Official CANSIM Documentation.

The Results

For those who are interested in seeing a walk through of the code, read through the results an head to “Data Science In R: Canadian Travel Data“.

First, let’s look at a table of the countries based on the number of times they appear in the top 15:

Country Frequency
China 15
Cuba 15
Dominican Republic 15
France 15
Germany 15
Italy 15
Mexico 15
Netherlands 15
Spain 15
United Kingdom 15
United States 15
Hong Kong 14
Ireland 12
Switzerland 11
Australia 6
Jamaica 5
Japan 5
Austria 4
Greece 2
Bahamas 1

Does anything here surprise you? This chart is lacking in much of the detail we need to draw conclusions. Instead, let’s take a look at the following Bump Chart. It ranks the countries as in order of number of visits over time. Note that the labels account only for the countries left in 2015, as some of the countries fall off the chart as time progresses.

BumpChartRecolor

There are some interesting take away from this chart. As probably expected, Canadians have always made the United States their number one travel destination. The United Kingdom was competing with Mexico for second place for a number of years before falling, along with France, to 5th and 6th place respectively in 2009. The Netherlands also sat comfortably at 10th place until nearly dropping off the chart in the past few years. What could be the causes of this?

Let’s try to quantify some of these trips. By using a stacked area plot on the number of trips (“Visits”) taken, we can get a sense of the relativity:

VisitsCount

Wow! The United States is overwhelmingly the largest importer of Canadian Travelers. Maybe not too surprising though, considering 75% of Canadians live within 100 miles of the American Border.

How Long Are We Staying?

For this, I’ve taken the average number of nights stayed in 2015 per visit. For interests sake I’ve also overlayed the cost of a flight to that country, chosen as the lowest result Google provides for a Wednesday flight six weeks from the writing of this article to the countries capital city. Note that for some countries, especially the United States, the cost of a flight can vary wildly, so these are far from rigorous.

AverageNights

Diving Into the US Data

Since the US represents such a large portion of the data, I thought it’d be interesting to take a deeper delve into the Canadian – American relationship. Based on the data available, we can answer a few questions:

How Many Visits Have Canadians Made to the US Over Time?

For this question, I’ve opted to show two ways of looking at the data. The unadjusted line shows the raw number of visits given by the data. I have also opted to make a population adjustment, bringing the number of visits to the 2015 level. This was done by using Statistics Canada Population Data, and using the following formula:

 Adjusted Visits 20XX = (Population of Canada 2015) / (Population of Canada 20XX) * (Number of Visits 20XX)

I have also added a couple notes which we can discuss after viewing the results:

Rplot01

You’ll notice a general upward trend, even adjusted for population, however there are some noticeable drops. I’ve made two suggestions to explain these drops:

  1. The 2008 Recession reduced the desire or ability to travel to the United States due to lack of funds
  2. The sudden drop of the $CDN from par ($1 CDN = $1 USD) around 2013 made it less appealing to travel to the US due to the cost.

Do you disagree or have additional thoughts? Leave me a comment and let me know.

How Much Are Canadians Spending While in the US?

This is a bit more tricky to answer. We have some competing forces at play here, and questions about how we calculate spending. Do we look at per night or per visit? How to we account for the exchange rate and inflation? Let’s first take a look at the raw data of average spending. Don’t forget that all the numbers are presented in Canadian Dollars.

AverageSpentUS

At face value, it looks like we have been spending more per night as time progresses. However, we can tease more information out of this data. Here is my methodology for adjusting the amount paid:

  1. Say, for example, we assume in 2002 we spent 85 Canadian Dollars
  2. The exchange rate in 2002 was about $1 USD = $1.5 CDN. Thus we spent 85/1.5 = $57 USD.
  3. Accounting for inflation at 2% per year, this is now worth $57 * (1.02)^13 = $74 USD in 2015.
  4. The exhange rate in 2015 was about $1 USD = $1.25 CDN. Thus we spent $74 * 1.25 = $92.50 CDN

To be a bit more technical, I have pulled the Canadian CPI to use as an inflation calculation. Additionally, the CAN-USD FX Rates are provided at January-1 of each year by Statistics Canada. Here is the resulting graph with these adjustments:

AdjustedSpent

This shows a much different picture. It looks like early in the decade, Canadians were spending more and more on their travels to the US, and in the more recent years they’ve been spending less. Take these results with a grain of salt – There are still many possible confounding variables and questions we could ask our self about the data, such as:

  1. Can we make an adjustment for the number of nights stayed?
    • One of the largest travel expenses for most people is likely the flight costs. The more nights you stay, the more nights this cost will be spread over.
  2. Are Canadians choosing different travel destinations within the US?
    • There are near countless vacation destinations within the US. These all come with their own costs and stay duration.

Come up with your own questions and have a try for yourself at answering some of them! You can find the code on my GitHub and a walkthrough of the code in my follow up article, “Data Science In R: Canadian Travel Data”

Sources:
Travel Data http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/arts37j-eng.htm
CPI: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/econ46a-eng.htm
Population: http://www5.statcan.gc.ca/cansim/a26
FX Rates: http://www.bankofcanada.ca/rates/exchange/legacy-noon-and-closing-rates/

Advertisements

One comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s