Passenger reviews reference architecture using big data lakes

摘要

Passenger reviews are quite popular in social media, e-commerce, and dedicated review websites. It is commonly referred as word of mouth which provides positive, neutral and negative sentiments from passengers about travel services. It helps other travelers to get insights through objective review ratings, subjective textual feedback and in cases along with media content like photos, audios, and videos. Processing passenger reviews is a key aspect for determining sentiments and gauging the pulse of passengers. Travel domain application systems are generally complex, diverse and typically managed by different stakeholders like airlines, airports, travel agencies, immigration, security and other services providers like cars, bus, trains, hotels, events. In this paper, we will propose a reference architecture using data lakes for managing passenger reviews where multiple stakeholders are involved along with challenges on volume, velocity, variety, and veracity. The end goal of the reference architecture is to provide a holistic approach to manage reviews with data gathering, processing and disseminating to a diverse set of stakeholders in the travel ecosystem for actionable insights.