Iten N. El Rouby

Assistant Professor

Tourism Department, Alexandria University, Faculty of Tourism and Hotels

 

 

ABSTRACT

The Egyptian revolution (25 January 2011) had a major impact on the Egyptian society, the economic sectors as well as a shift in the political relationships with the outside world. Tourism, as one of the major sectors upon which Egypt has been relying for the past years to cover deficiencies in the balance of payment has been seriously affected.

This study tends to use text mining applications to examine tourist sentiments towards Egypt in the period of 2011-2015 using samples of blog posts from a popular social media platform (TripAdvisor). Furthermore, the study applies correlation analyses to test whether there was a relationship between tourists’ sentiments towards Egypt at that period and actual tourist numbers visited Egypt at the same period.

The results of text mining analyses showed, that security problems played a major role in forming a negative sentiment towards visiting Egypt if also associated with foreign governmental bans to visit the country. These sentiments were clearly reflected in the number of tourists.

It is worth to mention, that also fluctuations between negative and positive sentiments and number of tourist visits were noted. That could be closely linked to different events occurring in the country at that period.

It is recommended that tourism stakeholders closely monitor social media platforms and provide accurate information to tourists to impact decision making.

Keywords: Egyptian revolution, Text mining, Tourism sector, TripAdvisor, Tourist visits.

 

1 INTRODUCTION

The Egyptian revolution, which had its outburst in the 25th of January 2011, was a result of anarchy and corruption that dominated the picture for several years. People flounced and the country's long-ruling regime was deposed.

The revolution and the numerous events that followed had a major effect on the tourism sector which is considered a highly sensitive sector if associated with political instabilities. Tourism is one of the important sectors that Egypt depends on to cover deficiencies in the balance of payment.

According to 2008/2009 statistics, the tourism sector covered about 42% of the deficiencies of the commercial balance of payment. This percentage fluctuated since then between increases and decreases till it reached 44% in 2012/2013(Ministry of Planning, 2014).

The objective of this research was to analyze the sentiments of actual and potential visitors in order to get useful insights about fears and concerns that overwhelmed visitors in the period between 2011 and 2015. The research also examined whether these sentiments could be linked to the number of tourists visiting the country at that period.

 

The research hyposized the following:

H1: Customer generated content are useful to extract valuable information about customer sentiments.

H2: There is a relationship between sentiment polarity of certain keywords used in reviews and number of tourist arrivals.

This research is divided into two parts. The first part includes some background information about an emerging field of research, namely sentiment and opinion mining and an overview of some events that occurred in Egypt after the revolution. This background information is necessary to understand the drop in tourist arrivals and the sentiments of visitors towards Egypt at that period. This part also covers a review of related literature and formulation of research question. The second part consists of a depiction of research approach and methodology that were used to test the hypotheses.

 

2 BACKGROUND ON TEXT MINING AND SENTIMENT ANALYSIS

Sentiment analysis or opinion mining refers to “the application of natural language processing, computational linguistics, and text analytics to identify and extract subjective information in source materials” (Katarzyna et al., n.d.)

In today`s era of Internet technologies and associated social media avenues such as blogs, discussion forums, peer-to-peer networks and other types of social media, consumers have unprecedented possibilities to share their experiences and opinions regarding any product or service. Businesses and destinations have nowadays exceptional opportunities for information retrieval. This huge number of consumer reviews offers a chance for businesses to analyze and make statistical inferences about consumer behavioural patterns in the field of tourism and hospitality.

In late 2012, Expedia’s set of verified reviews reached a total number of more than7.5 million (Media room Expedia, 2012). This huge number of available customer generated content could be effectively used to develop business intelligence strategies. These reviews reflect actual experiences and generate word-of-mouth marketing for certain brands or destinations. If not carefully monitored, this indirect marketing opportunity could harm brand reputation and destination images. Reputation management strategies are nowadays widely used to limit the negative word-of-mouth that can be quickly spread through the viral effect of social media.

The challenge that faces businesses and destinations is how to convert this wide range of data into useful insights. There are several approaches that can be used to extract useful information from social media such as channel reporting tools, overview score- carding systems and predictive analytic techniques like text mining (Pang et al., 2008).

Companies and destinations are becoming more convinced, that the viral effect of social media can play a major role in shaping the opinions of other consumers, their brand loyalties and their purchase decisions. Sentiment-analysis technologies can be effectively used for extracting opinions from unstructured human-authored documents. This could help travel businesses and destinations change their marketing messages, consider different tools for brand building or develop different strategies for positioning.

Nevertheless, this data can be used to draw general insights about overall sentiments and opinions but need further investigation to stand upon customer needs, behaviours and preferences (Pang et al., 2008).

The objective of this research was to extract useful insights about tourists` sentiments from online customer generated content posted on TripAdvisor. The reviews could be categorized as being posted by tourists that visited the country and were sharing their experience. Another category of reviews consisted of inquiries and investigations from those under consideration for a future visit. These potential visitors were seeking advice and were exposing their fears to go through with their plan to visit the country in the period of instabilities from 2011till 2015.

 

3 AN OVERVIEW OF THE POLITICAL SITUATION IN THE PERIOD OF 2011-2015 (MC-DAOULYIA, 2015).

The following events constitute the most important stations in the history of Egypt since the date of (January 25) until today.

On January 25th, more than a million demonstrators gathered in Cairo's Tahrir Square, which has become a symbol of the revolution. Demonstrations continued in Tahrir Square in Cairo and other parts of Egypt and the number of protesters escalated as days went by.

On February 11, 2011 Mubarak, after several speeches, stepped down and a military council headed by Supreme Commander of the Armed Forces announced running the country on a temporary basis.

In January, 2012 the Supreme Judicial Elections Commission announced that the Islamists have received more than two-thirds of the seats in the parliamentary elections.

June, 2012 Morsi won the run-off in the presidential election by 51.7%, and led the country on the thirtieth of June to became the first president of Egypt after free elections.

Due to chaos that ruled the country, a campaign to collect signatures calling for the isolation of Morsi and the holding of early presidential elections was launched.

In June 2013, on the first anniversary of Morsi`s presidency take-over, millions of Egyptians began mass demonstrations for days to demand him stepping down.

In July 3rd 2013, Sisi announced the isolation of Morsi and the inauguration of the President of the Constitutional Court Counselor temporarily as president. In May 2014, Sisi won presidential election by 96.91 percent.

All these events had major effects on Egypt internally and externally. The economy of Egypt in particular was majorly influenced by the events succeeding the revolution. Tourism, as a sensitive industry was extremely affected by the political unrest that ruled the country as will be presented later in succeeding parts of this research.

 

4. RELATED WORK AND RESEARCH QUESTION

With the emergence of social media and associated UGC a new stream of research tackled the approaches of sentiment analysis and opinion mining using language processing techniques (Hu et al., 2004; Pang et al., 2008). Some work was also devoted to automated extraction of product reviews (Lee et al., 2011).

Other researchers studied the affect of product reviews on product sales with the focus on numeric review ratings (e.g., Godes et al., 2004; Chevalier et al., 2006; Liu 2006; Dellarocas et al., 2007; Duan et al., 2008; Forman et al. 2008).

Decker et al. (2010) used text mining to forecast the effect of product features and brand names on the overall evaluation of the products. A research by Ghose et al. (2012) aimed at improving the recommendation strategy for travel search engines in order to provide customers with most suitable hotel choice early on the search process.

A study by Marrese-Taylor et al. (2013) determined consumer preferences about tourism products, particularly hotels and restaurants, using opinions conveyed in customer reviews on TripAdvisor. Results showed that tourism product reviews could provide precious information about customer preferences that can be extracted using aspect-based opinion mining approaches.

The purpose of the study by Claster et al. (2013) was to examine whether tweets could convey market intelligence opportunities. The results showed that tweets or micro-blogs if sentiment mined can be used as a valuable source of information for market research in the tourism and hospitality industry.

A study by Banerjee et al. (2015) examined to what extent authentic and fake reviews could be distinguished. The researchers used supervised learning algorithms based on four linguistic clues, namely, understandability, level of details, writing style, and cognition indicators for analysis.

A research by ComScore (2007) revealed that customer generated content has a more significant influence on customer sales if compared to content generated by professionals.

The study by Xiang et al. (2010) confirmed the growing significance of social media in e-tourism and the importance of them as a source of information.

 

Based on the review of the literature the research question arised:

Could customer generated content be used to extract valuable knowledge about customer preferences and behavioural patterns?

In order to answer this question, the study adopted a methodology that extracted data from TripAdvisor reviews to examine sentiments and opinions of actual and potential visitors towards Egypt in the period of 2011-2015. The research also went beyond overall text polarity for sentiment analysis and shed light on some keywords that expressed concern and fears in order to get more subtle insights about customer needs, concerns and worries.

 

5 DATASET AND METHODOLOGY

TripAdvisor, a popular social media platform, was founded in 2000 and currently covers more than 4.9 million accommodations, restaurants, and attractions. TripAdvisor contains 225 million travel reviews and opinions written by 5 million registered members and counts 340 million visitors per month. People on TripAdvisor can exchange information about destinations, tourism products, services, travel experiences, weather, shopping or any other topic (TripAdvisor, 2015).

In order to conduct sentiment analysis for TripAdvisor reviews, the researcher relied on an open source text mining software called Semantria. This software accomplishes language processing tasks. It includes several features like overall text sentiment analysis, queries, intentions, facets, entities, language among others. Some of these features can be customized by the user.

The data collection phase entailed gathering reviews from TripAdvisor covering the period of 2011-2015. The phrase “Safe to travel to Egypt” was used in TripAdvisor search engine. The query generated a total of 2352 reviews. The results were filtered (data cleaning phase) and a total of 895 unrepeated reviews were extracted for the purpose of this research. The software is an excel add-in, so the results of the analysis were presented in the form of an excel sheet accompanied by some charts of selected features. The analysis focused on features such as overall text polarity (sentiment), queries, entities...etc.

 

5.1 Results

The data set of 2011:

Semantria uses an algorithm to calculate sentiment polarity. The document sentiment has a range spread from -2 to 2, where -2 is really negative, -1 is negative, 1 is positive, and 2 is really positive.

The data set of 2011, which comprised of 101 reviews, showed that the overall polarity of the reviews was positive (84%) while 4% were negative reviews and 12% were neutral reviews (Table1).

Table 1: Overall polarity of reviews according to year.

Source: Semantria, 2015.

 

This study tended to examine the effect of the Egyptian revolution on the sentiments of actual tourists and also on potential visitors. In order to examine this phenomena, special keywords related to “safety”, “security”, “curfew”, “traffic”, “ politics”, “service”…etc were chosen and set up in the query feature. The query count showed that in 2011 the keywords related to category “safety” appeared 100 times in the reviews and had an overall neutral polarity. The word “service” appeared 34 times with neutral polarity. The word “politics” appeared 31 with also neutral polarity (Table 2).

 

Table 2: Query, Query Count, Query Sentiment Score and Query Sentiment Polarity of TripAdvisor reviews of 2011

<img src="data:<;base64,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