Team assignment

Delta Airlines $$ Sales Forecasting

It is the summer of $2021.$ You are working as a financial analyst at Delta Airlines. After years of steady growth, the Covid$-19$ pandemic, lock downs and travel bans have wreaked havoc in the airline industry. In the early days of the pandemic in $2020,$ air travel almost came to a standstill. Many airlines responded by downsizing, laying off pilots and staff, and temporarily decommissioning planes to manage cash burn. Fast forward a year later, domestic demand has redounded and is on the path towards recovery after lockdowns and some travel restrictions have been lifted.

In this time of uncertainty, forecasting demand is of vital importance as management needs to create a financial plan for the coming third quarter, historically the strongest in the airline industry when many passengers undertake their summer travel. An accurate forecast is essential for business planning as it determines hiring, scheduling of crews, planes and routes, and enables management to issue an earnings forecast for the third quarter to the investment community.

Management has asked you to deliver a data$-$driven forecast of domestic passenger demand for the third quarter of $2021.$

$1)$ Plot the time series of domestic ticket $($#passengers$)$ across time. What elements of a forecast do you notice:

$$ A trend?

$$ Seasonality?

$$ Disruptive events?

If present, describe how do they manifest or do not manifest in the pattern your time series plot reveals.

$2)$ Almost every time series sales data follows a seasonal pattern. Compute a seasonality index per quarter and de$-$seasonalize your time series.

$3)$ Implement three different time series models using the entire time series on passenger demand.

$$ Moving average forecast with a four quarter rolling forecast window

$$ Exponential smoothing with $\backslash$alpha of your choice

$$ Double exponential smoothing with trend adjustment $(\backslash$alpha $,\backslash$beta of your choice$)$

Hint: Don$$t forget to re$-$seasonalize your forecast.

$4)$ Assess the general quality of your forecasts in fitting the time series for each of your three forecasts.

$$ Report the mean absolute percentage error.

$$ Plot the actual and forecasted sales in a time series plot.

$5)$ What sales forecast does each model provide for the third quarter of $2021?$

$6)$ The exponential smoothing forecasts you implemented rely on parameter choices such as selecting $\backslash$alpha $,\backslash$beta $,$ and will determine what forecast the model produces for the future third quarter of $2021.$ Under which circumstances $($time series properties$)$ should you choose $\backslash$alpha $,\backslash$beta close to $1$ or close to $0,$ respectively? Which one of these choices seems more appropriate to forecast third quarter ticket demand for Delta?