Airline-Traffic-Predictions

After applying for a Data Scientist role at Qantas, they requested applicants complete a brief data science task to advance to the next stage of the interview process.

The task description and process of evaluation were:

Excercise

The Bureau of Infrastructure, Transport and Regional Economics (BITRE) publishes monthly air travel statistics between various cities. The task is to build a model of air passenger traffic from Sydney to Los Angeles using this data. This model should predict the volume of passengers for each month in 2019.

When building this model, please keep in mind the following:

• We are only concerned with passenger traffic.

• We are only concerned with outbound traffic from Sydney to Los Angeles and not the return trip (Los Angeles to Sydney).

• Air traffic can be highly seasonal in nature and comparisons are often made in comparison with a similar period in the previous (or next) year.

The data can be obtained from the link below. The specific dataset you should use is titled “City pairs data – passengers, freight and mail – 2009 to current”; the data can be found in the ‘Data’ tab of this spreadsheet. No other data source is required.

https://bitre.gov.au/publications/ongoing/international_airline_activity-time_series.aspx

The solution should:

1. Load the source data. You may copy the relevant data to a CSV file to simplify the loading. You are not required to supply this data as part of your solution.

2. Build a model that is appropriate for these data.

3. Use the model to forecast the air passenger traffic from Sydney to Los Angeles for each month in 2019 and print this forecast to the console (stdout).

The Solution

You may spend as much time as you need on this exercise, but we would prefer you to not spend more than 1-2 hours.

The deliverable for this exercise is a single source file (.py or equivalent for other languages) or a Zip file if your solution has multiple source files.

We will be assessing your solution by:

• Your ability to interpret and manipulate data;

• Your ability to apply a principled and mathematically rigorous approach to solving a data science problem; and

• Your ability to write clear, concise and readable code.

When working on your solution, please keep in mind that:

• We are not looking for one “right” solution. Rather, we are looking for you to demonstrate your ability to choose a reasonable approach amongst the many possible approaches.

• We will not be benchmarking your model against a ‘ground truth’.

• We are not looking for extensive validation of your results.