Progress in data science means it is now possible to predict when business travellers will cancel hotel bookings, as well as identify other key issues.
Travellers’ data can now be analysed with a great degree of certainty thanks to sophisticated modelling, an expert revealed at the Business Travel Show in London.
Daniel Reeve, Capita Travel and Events chief technology officer, said previous estimates of there being 1.7mb of data collected on every person every second of the day was probably an underestimate.
He said when it came to business travel, management information, business intelligence, data analysis and data science were key.
Reeve, however, said one drawback with management information was its tendency to be “after the fact reporting”, while business intelligence permitted only occasional forecasting.
In contrast, he said data science captured the information it needed and was predictive, via the use of statistical modelling and machine learning. “It works out what information will be most helpful,” said Reeve.
Computer scientists “with a strong business sense” developed the technique, Reeve said. “They are smart cookies,” he told delegates at the show on Wednesday (26 February).
He gave an example of how Capita used two hotel chains and three different data science models to predict room cancellations for a client. They varied between 83% and 95% accuracy.
“So how can we use that? We can target potential guilty travellers, but also identify others in advance of [the hotel’s] cancellation policy so rooms can be cancelled without penalty.”
Such information, he said, could also be very useful to hotel brands.
Data science was also used to further “smart working”. Another study found there was no increase in employee absenteeism after trips that involved travelling at weekends or overnight and for short trips versus longer stays.
However, there was a significant difference when longer working days were added to the travel stress. Data science was used to produce models that were 96.7% to 97.3% accurate in their predictions, said Reeve.