Thanks to the Big Data processing developed by Mabrian we can assess visitors’ satisfaction of a specific destination. This index, called Global Tourism Satisfaction index, is elaborated by crossing data in relation to the perception of the tourism product – calculated on Mabrian’s TPi-, the sensations reported about the area’s security –Mabrian PSi- and the satisfaction indicator regarding climate – Mabrian’s PCi.
In some places we also elaborate the hotel satisfaction index (HSi), in which case the global satisfaction index can be crossed with the HSi to obtain different perspectives.
Mabrian indexes help in the planning and implementation of tourism strategies and can be used from tourist institutions and agents to investment companies. They are based on real data which is collected in real time and they show fluctuations in the visitor’s perceptions throughout the year, as we can see on the following graphs, related to the island of Menorca:
In this case, the Global index shows a clear difference between the destination on high season (August 2017) and low season (February 2017). The reason for the more than 9 points difference between one and the other is found on the variation between climate satisfaction indexes and tourism satisfaction indexes. This last one clearly conditions the visitor’s opinions about Menorca during the low season: the most marked ballast during February is in the tourist offer, more than in meteorological circumstances.
Each one of the indexes is built from data offered by various platforms through their API (Application Programming Interface). There is also data obtained from other sources, like global distribution systems (GDS), and with the capture and processing of information offered by specific websites.
For all this flow of data to copper the analytical value that we want to bring, it is filtered with a preprocessing. This includes a set of algorithms with functional rules. In this stage, data from bots and other automatisms is discarded, and the information acquired is verified to come from people visiting the destination and not from habitual residents.
The preprocessing also incorporates a specific word filter built by empirical experimentation that admits or discards mentions based on their contextual and historical load.
The information is finally filtered with a classic learning algorithm (spam vs. ham) that detects mentions referred to tourism. In this phase, the sieve incorporates criteria linked to current news.
All data is treated with Natural Language Processing (NLP) in a multi-language format that includes segmentation of sentences, tokenization, stemming (reduction of the word to its root), grammatical tag, etc.
Each one of the collected and processed data segments go on to determine their corresponding index through its own equation. It is an algorithm configured from our experience, centered and specialized in tourism Big Data. Mabrian indexes are a powerful analysis and prospecting tool that monitor and forecast behaviour in specific markets and can be used to make the most appropriate decisions on each moment.
