Model 3: Association

It all started when…

Association could reveal relationship of multiple factors. If knowing factor A oftern appears together with customer satisfaction, then we could derive some recommendations about increasing the factor A to increase the possibility of customer satisfaction.

In this algorithm, we don't separate customers into different groups. Instead, we put the cusomter information in the model to check whether demography really matters. All ratings are integers. To interprete the result easier in this part, we define 3 levels of ratings:

'high': rating = 5

'mid': rating = 3 or 4

'low': rating = 1 or 2

Top 10 rules involving 'satisfied flight experience'

Top 10 rules involving 'dissatisfied flight experience'

Conclusion

From top association rules, we could find that people who are satisfied with their flight are usually of business class or having business travel. 'loyal customer' also appears a lot in the antecedents, but being a member of the flight company has reciprocal causation with being satisfied with their experience.

For the dissatisfied customers, we could find that these people are usually of eco class or having short travel. From top 5 rules, we suggest airline companies to improve online wifi service, gate location and time convenience of the flight.

We don't find any demographic variables in those top rules, so maybe they are not the keys impacting flight satisfaction

——————-——————-——————-——————-——————-——————-——————-——————-——————-——————-——————-——————-

For all groups of customers, improving the wifi service could improve customer satisfaction.

For long-flight customers, improving inflight entertainment would be the most effective method to improve customer satisfaction.

For ecoplus-class customers and eco-class customers, improving seat comfort would be the most effective method to improve customer satisfaction.