Gulf Air introducing a novel way for airlines to reach out to their travellers and find out what makes them tick
The importance of social media in a modern business cannot be understated. Serving as an invaluable tool in driving brand recognition and connecting with their target market, many businesses today utilise social media in some form or another.
What is perhaps more uncommon however, especially here in the Middle East, is the analysis of social media in Arabic, as a means of devising a successful business strategy.
This was the step undertaken by Gulf Air, the national carrier of the Kingdom of Bahrain. One of the first commercial airlines set up in the Middle East back in 1950, Gulf Air is today one of the most technologically advanced organisations active within the region’s aviation market.
Because of the rising popularity of social media, the team at Gulf Air sought to create a system that not only analysed social media posts in both Arabic and English, but was sophisticated enough to produce insights on consumer satisfaction and behaviour.
While such insights could be produced by monitoring social media manually, it is an incredibly daunting and time-consuming task, and one that is prone to human-error.
The answer to the problem was a sentiment analysis platform based on big data technologies.
“As an airline, it is crucial to know what our customers are thinking. We want to know who is satisfied with Gulf Air’s service, who is not, if not then why, and what Gulf Air can do to improve its offerings,” says Dr. Jassim Haji, director of information technology, Gulf Air.
“Lots of our passengers as you could imagine are Arabs and they speak in Arabic and they tweet in Arabic,” comments Dr. Haji.
“No one uses this data in any meaningful way, except listening to jokes or complaints… So we came up with an idea of ‘why don’t we look into this data that is available and analyse it and put it into some structure?’ Thus Arabic Sentiment Analysis was born.”
Based upon Cloudera’s distribution of the open-source Hadoop big data framework, Arabic Sentiment Analysis was developed over a nine-month period and fully in-house by a team comprised of just six members.
With the system already up and running, Arabic Sentiment Analysis currently processes thousands of social media posts, analysing and providing results on what passengers are discussing, as well as their feelings on particular routes, products and services.
Furthermore, the system provides Gulf Air a wider analysis of the state of the market, and sometimes even of the actions undertaken by their competitors.
“Of course we needed a huge database of Arabic words,” comments Dr. Haji on the scale of the project.