Air Travel Demand Forecasting: Before and After Covid-19

It has been repeatedly emphasised that the COVID-19 epidemic and its effects on the airline sector are unparalleled. With air travel demand forecasting transitioning from more traditional methodologies based on “gut feeling” and “rules of thumb” to more scientific approaches based on data, Artificial Intelligence (AI) models were just beginning to find their place in airline planning processes when COVID-19 struck and levelled the playing field. AI models, which frequently rely on vast amounts of data to provide precise forecasts, have unexpectedly lost access to their most important source of fuel: past data.

An environment is deemed uncertain in artificial intelligence (“AI”) “if it is not entirely observable or deterministic.” Although one of the key goals of AI is to enable humans to “make better decisions under uncertainty,” the question that needs to be asked right now is if COVID-19 has increased levels of uncertainty in aviation that are too high for precise estimates of travel demand. After all, this new world is not accurately represented by previous facts. What really can we anticipate from AI models in terms of prediction accuracy and horizon airlines, given the various events and constraints that affect air travel demand forecasting occurring daily?

Before COVID 

Prior to COVID, airlines gradually began to understand the value of data and digitization. Processes for network planning, scheduling, and fleet assignment relied more and more on past travel trends to forecast future travel demand. Significant increases in profitability were made possible by real-time cost data feeds into airline network planning.

Previously, external measures like economic activity were utilised to inform decisions about establishing new origin-destination (OD) pairs. However, even before COVID-19 was interrupted, it was noted that such indicators were insufficient for accurately forecasting future demand. This is especially true considering that demand should be estimated across all modes of transportation for an OD pair (air, rail, etc.), along with passenger profiling, to understand their modal preferences better. When the aviation sector realised this, it began to embrace data science, just in time for the epidemic to suffocate it.

After COVID

The need for the aviation sector to employ a variety of data sources to understand better future demand, including both short- and long-term changes, is perhaps more important than ever. AI demand forecasting models should use data from online travel agency searches, social media buzz, event planning, professional networks, etc. Additionally, anonymized data from sources like credit card data, social media, etc., should be added to them in order to understand better the choices made by travellers when using various forms of transportation.

In particular, data on government travel regulations, current events, data on working from home, and possibly even COVID-19 spread forecasting future demand models need to be dynamically infused into demand forecasting models in order for them to be able to respond and adjust to shocks like the COVID-19 pandemic. These sources could all be infused in addition to a number of other data sources that need to be examined through correlational analysis.

It is crucial to create indicators that could measure how events throughout the world affect travel patterns so that airlines can modify their itineraries as necessary. This will be even more important in the future, when we may anticipate regular and erratic postponements, reschedules, and cancellations of those activities.

Additionally, it has been demonstrated that markets can be grouped based on specific features, with the resulting data being fed back into demand forecast models. The models could be modified utilising the data gathered by additionally clustering markets according to different pandemic indicators as the epidemic develops at various scales around the world. Airlines might organise markets into clusters of like locations in which the travel demand is anticipated to behave similarly by relying on unsupervised machine learning techniques, which would minimise the requirement for large amounts of historical data.

So, if you are also looking for the right forecasting future demand solution, then RateGain is your place to be. Connect with them now and get the best air travel demand forecasting solutions today.