Time Series Forecasting for Business Plans

Forecasting is always an interesting topic.  We could precisely calculate the trajectory path of a free fall mass using the law of physics but we could not predict a thunderstorm accurately in two weeks’ time.  Yet, forecasting remains as a very important practice in many areas such as business, economics, medicine, transportation, politics and others.  It helps to improve business and resource planning.  Over the years, it has attracted many researchers to develop various forecasting methodologies.

By studying the past observations of a time series data, and then develops a model that fit the pattern of the past data is called time series modelling.  Once the model has been developed, it could be used to forecast the near future behaviour of the underlying process.  One of the common methodologies is the Box-Jenkins method, or sometimes known as Autoregressive Integrated Moving Average (ARIMA) model.

Diagram 1 shows the actual vs forecasted data of Kuala Lumpur International Airport (KLIA) international passengers traffic from Jan 2012 to Dec 2015.  The red curve is the actual traffic data taken from Malaysia Airport Holding Berhad (MAHB) annual report (Read more here).   Three different traffic scenarios – typical, optimistic and pessimistic, were forecasted using ARIMA and were plotted as blue, green and orange curves respectively.  In Diagram 1, the pessimistic scenario trended closely to the actual data.

Diagram 1.  KLIA International Passengers Traffic Jan 2012 – Dec 2015

Diagram 2 shows the actual vs forecasted data of KLIA international passengers traffic from Jan 2012 to Dec 2016.  The legend labelling is the same as in Diagram 1.  In this case, the typical scenario trended similarly to actual data.

Diagram 2.  KLIA International Passengers Traffic Jan 2012 – Dec 2016

Time series forecasting methodology such as ARIMA is a very useful tool for companies to manage risk or project budgeting in a systematic way.  The forecasted results provide an insight to the magnitude of the unforeseen circumstances.  As such, the management could react to future developments promptly and confidently, having developed at least three possible outcomes.

For more information about time series forecasting and projections in setting business plans, please visit http://www.mpcap.com.my/ or contact info@mpcap.com.my.