{ "paper_id": "2021", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T07:33:24.318297Z" }, "title": "User Generated Content and Engagement Analysis in Social Media case of Algerian Brands", "authors": [ { "first": "Aicha", "middle": [], "last": "Chorana", "suffix": "", "affiliation": { "laboratory": "Laboratoire d'informatique et Math\u00e9matiques", "institution": "Universit\u00e9 Amar Telidji Laghouat", "location": { "country": "Alg\u00e9rie" } }, "email": "a.chorana@lagh-univ.dz" }, { "first": "Hadda", "middle": [], "last": "Cherroun", "suffix": "", "affiliation": { "laboratory": "Laboratoire d'informatique et Math\u00e9matiques Universit\u00e9 Amar Telidji Laghouat", "institution": "", "location": { "country": "Alg\u00e9rie" } }, "email": "hadda_cherroun@lagh-univ.dz" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "Nowadays, online social media hugely influences individuals' daily lives, companies, institutions, and governments. Analyzing the online social content related to the productivity of any company becomes crucial to manage and supervise its activities and future trends. We investigate the quality of social signals and content related to Algerian products and services to enhance their exploitation and deployment. Our investigation relies on the statistical analysis of social signals and the textual analysis of User-Generated Contents (Posts and Comments). The current work has been done on a sample of more than 50 brands gathering products and services on Facebook with 10K posts and their related comments totaling around 100K. We measure Users/Brand Engagement Rates (ER) considering reactions and content. We adopted a statistical analysis for the reactionbased measurement. We leveraged an LDAbased Topic Modeling Approach for contentbased measurement. Our findings emphasize the significance of the existing social signals and user-generated content in the Algerian context.", "pdf_parse": { "paper_id": "2021", "_pdf_hash": "", "abstract": [ { "text": "Nowadays, online social media hugely influences individuals' daily lives, companies, institutions, and governments. Analyzing the online social content related to the productivity of any company becomes crucial to manage and supervise its activities and future trends. We investigate the quality of social signals and content related to Algerian products and services to enhance their exploitation and deployment. Our investigation relies on the statistical analysis of social signals and the textual analysis of User-Generated Contents (Posts and Comments). The current work has been done on a sample of more than 50 brands gathering products and services on Facebook with 10K posts and their related comments totaling around 100K. We measure Users/Brand Engagement Rates (ER) considering reactions and content. We adopted a statistical analysis for the reactionbased measurement. We leveraged an LDAbased Topic Modeling Approach for contentbased measurement. Our findings emphasize the significance of the existing social signals and user-generated content in the Algerian context.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "Several companies harness the potential of Online Social Networks (OSN). OSN present an effective communication channel between the company and its customers (Anubha and Shome, 2021; Santoso et al., 2020; Voorveld, 2019) . Indeed, these social networks, tremendously, scale up the network effect of standard marketing techniques such as Word-Of-Mouth. Thereby, the emergence of Social Media Marketing(SMM). Indeed, SMM has become an independent field of marketing for which many opportunities have been recognized: i) raising public awareness about companies, ii) product development through community involvement, by analyzing User-Generated Content(UGC) and gathering experience for the future steps (Richter et al., 2011) .", "cite_spans": [ { "start": 158, "end": 182, "text": "(Anubha and Shome, 2021;", "ref_id": "BIBREF3" }, { "start": 183, "end": 204, "text": "Santoso et al., 2020;", "ref_id": "BIBREF16" }, { "start": 205, "end": 220, "text": "Voorveld, 2019)", "ref_id": "BIBREF18" }, { "start": 702, "end": 724, "text": "(Richter et al., 2011)", "ref_id": "BIBREF15" } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "The analysis of UGC in social networks has fundamentally reshaped marketing strategies. Users have unlimited freedom to express their opinions through different interactions (e.g. reviews, like, rating. . . ) on web resources. This rich source of social information can be analyzed and exploited to serve several applications in various contexts. In particular, opinion mining and sentiment analysis techniques that have the ability to reveal users' behavior or reaction regarding an item or event. This knowledge represents the bedrock to build an effective content-based recommender system (Zatout et al., 2019) .", "cite_spans": [ { "start": 592, "end": 613, "text": "(Zatout et al., 2019)", "ref_id": "BIBREF20" } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Users/Brand owners' Engagement analysis and measurement in Arab-world companies seem to be falling behind and show somewhat shy usage. This paper investigates the existence and magnitude of social Media Marketing and explores the nature of both companies' and users' engagement. We also focus on the analysis of textual User-Generated Content in order to present some of their salient features by answering the following questions:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "\u2022 Are there enough social data on Algerian productivity that can be harnessed to improve Recommender systems applications?", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "\u2022 What are the most used social signals? \u2022 How are the users engaging in Social Media Marketing?", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "\u2022 Is social data quality significant to build learning models? Such as Ranking Algerian products, Predicting some economic phenomenon, etc.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "The rest of this paper is structured as follows. In the next section, we present some background on Social Signals, concepts of Brandcommunities and brand-owners engagement, and how they can be measured. In addition, we review some related work. In Section 3, we describe the followed process in this investigation, starting from the targeted sample of data to the data analytics step. Section 4 is dedicated to reporting results and findings with discussion. We conclude in Section 5.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "In this section we give some preliminaries on the engagement of brand-owners and their brand-communities (users) through social signals and how this engagement can be measured. Then, some related work are discussed.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "Engagement in social media, is a multifaceted complex phenomenon that can be measured by a number of potential approaches (Lalmas et al., 2014; An and Weber, 2018) : i) Self-Reporting Approaches ii) Physiological Approaches and iii) Web Analytic Approaches. This latter refers to the extraction of parameters thought to influence users' engagement, from the digital traces (UGC)left by users while interacting with a website. The most popular UGC on the Web are social signals such as comment, tag, Emotion, Post Message, Reaction, Share, vote, etc Most of these signals are mainly introduced to enable users to express whether they support, recommend or dislike a content (text, image, video, etc.) . We can distinguish between social activities' actions and reactions. The actions (e.g., like, share) with counters indicate the rate of interaction with the Web resource. While the reactions, introduced last years, are emotional signals that allow users to interact with posts in a quick way using one of the reactions(Like, Love, Haha, Wow, Sad, and Angry) to react even if the content is difficult to like, as in the case of gloomy news.", "cite_spans": [ { "start": 122, "end": 143, "text": "(Lalmas et al., 2014;", "ref_id": "BIBREF8" }, { "start": 144, "end": 163, "text": "An and Weber, 2018)", "ref_id": "BIBREF2" }, { "start": 673, "end": 699, "text": "(text, image, video, etc.)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "Concerning the metrics, for i) Brand Engagement, we consider the metrics related to brand's posts: Content and Media Type and their related users interactions. While for ii) User Engagement, the considered metrics are: Reaction rates, the relevance of textual generated content regarding the related Brand/service.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "Considering the scarcity of investigations on measuring Brand/User engagement for the Algerian Brands, we have narrowed our literature review to some related work from the Western world (Pletikosa Cvijikj and Michahelles, 2013a; Jayasingh and Venkatesh, 2015; Olczak Concerning related work from a Natural Language Processing (NLP) point of view, we can consider that there is a lack of statistical and content analysis of social signals in the Algerian context. For that, we restricted our review to some Algerian online content corpora built for the purpose of content-based analysis, mainly opinion mining and sentiment analysis.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "For the sentiment analysis purpose, Mataoui et al. (Mataoui et al., 2016) have built a dataset for Algerian dialect from some main frequented Algerian pages. The chosen social signals are textual (text of posts and comments). They have annotated the dataset manually and they built three Algerian lexicons.", "cite_spans": [ { "start": 51, "end": 73, "text": "(Mataoui et al., 2016)", "ref_id": "BIBREF9" } ], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "Rahab et al. (Rahab et al., 2017) have built ARAACON (ARAbic Algerian Corpus for Opinion Mining), a corpus of comments collected from online Algerian Arabic journals. These comments are mostly written in Algerian Dialect.", "cite_spans": [ { "start": 13, "end": 33, "text": "(Rahab et al., 2017)", "ref_id": "BIBREF14" } ], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "From an economic side, recently, some studies, have investigated the impact of social media on digital businesses. For instance, Graa et al. (Graa et al., 2017) have studied the impact of social media on Algerian purchase behavior. While (Abuljadail and Ha, 2019) have studied the impact of post content type (Hedonic and utilitarian benefits) on the engagement rate. However, these studies are done by means of traditional questionnaire surveys.", "cite_spans": [ { "start": 141, "end": 160, "text": "(Graa et al., 2017)", "ref_id": "BIBREF6" }, { "start": 238, "end": 263, "text": "(Abuljadail and Ha, 2019)", "ref_id": "BIBREF0" } ], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "In (Soumeur et al., 2018) , authors have focused on the specificity of Algerian dialect. They performed a specific pre-processing that improved the data quality. In order to perform sentiment analysis, they used two machine learning models: a Multilayer Layer Perceptron (MLP) neural network and a (Deep) Convolutional Neural Network (CNN).", "cite_spans": [ { "start": 3, "end": 25, "text": "(Soumeur et al., 2018)", "ref_id": "BIBREF17" } ], "ref_spans": [], "eq_spans": [], "section": "Background and Related Work", "sec_num": "2" }, { "text": "In this section, we present an overview of the followed steps, as illustrated in Figure 1 . We start by data collection, followed by data preparation (annotation and pre-processing), then data analytics by means of some measured aspects.", "cite_spans": [], "ref_spans": [ { "start": 81, "end": 89, "text": "Figure 1", "ref_id": null } ], "eq_spans": [], "section": "Methodology", "sec_num": "3" }, { "text": "Considering the scarcity of datasets on Algerian social signals related to brands and their communities. We have been constrained to collect a sample that encompasses the most powerful and well known brands/services and industrial companies in Algeria. In addition to their visibility on Social Media. The dataset categorizes the collected Brands and Services according to their topic of interest.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data collection", "sec_num": "3.1" }, { "text": "Following a similar recipe to the one suggested by authors in (Bougrine et al., 2017) . The sample dataset has been collected by following these stages:", "cite_spans": [ { "start": 62, "end": 85, "text": "(Bougrine et al., 2017)", "ref_id": "BIBREF5" } ], "ref_spans": [], "eq_spans": [], "section": "Data collection", "sec_num": "3.1" }, { "text": "Inventory of Potential Algerian Brands/Services", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data collection", "sec_num": "3.1" }, { "text": "Brand-Owner Generated Content Analysis", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Collection", "sec_num": null }, { "text": "Post Characteristics", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Analytics", "sec_num": null }, { "text": "User-Generated Content Analysis ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Measured Aspects", "sec_num": null }, { "text": "We have prepared the data following two step, namely, annotation and cleaning. We manually annotated users' comments according to their:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "\u2022 Relevance : we have considered two classes. Relevant: that says that the topic comments have relation with the targeted post and Irrelevant: which does not have any relation with the related post.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "\u2022 Polarity : Positive, Negative or Neutral.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "\u2022 Language distribution and used scripts:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "We have considered the most used languages for the Algerian community which are Modern Standard Arabic (MSA), the first and second foreign languages (French and English), and the Algerian Dialect as the common communicated language in the community. In fact, for each comment, we considered the ratio of words by language.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "For the purpose of the textual content analysis, we adopted the following data-cleaning steps for all comments in our dataset. First, we remove all photos, stickers, and punctuations, keeping only textual data. Then, we remove stop words (Arabic and French stop words).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "After that we apply tokenization. We also remove emojis in a second round of cleaning the data.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Annotation & Cleaning", "sec_num": "3.2" }, { "text": "In order to investigate the nature and rates of both users and brands' owners engagements, we adopted two types of analysis considering User Generated Content UGC and Brand Generated Content BGC respectively. In what follows, we demonstrate the considered metrics for both types.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Analytics & Measured Aspects", "sec_num": "3.3" }, { "text": "We addressed user engagement in two ways. One relies on statistical reaction-based analysis, where Engagement Rates consider simple metrics like the number of shares, comments, and reactions) (Pletikosa Cvijikj and Michahelles, 2013b; Perreault and Mosconi, 2018).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "The second metric relies on content analysis (linguistic features, comments' text analysis) where we deploy some (NLP) techniques to measure the quality and rate of the engagement. These techniques include applying Topic Modeling on comments using Latent Dirichlet Allocation model(LDA) 4.3) (Blei et al., 2003) . Furthermore, we measure the user engagement rate based on content analysis for post/brand using the relevance of users' comments regarding the post content. Thus, we suggest the following formulas:", "cite_spans": [ { "start": 292, "end": 311, "text": "(Blei et al., 2003)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "First, the content-based engagement rate with a specific post CP ER (1) metric.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "CP ER = RCP N CP", "eq_num": "(1)" } ], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "Where RCP and N CP are the number of relevant comments per post and the total number of comments per post, respectively. We rely on Topic Modeling of comments' to achieve such goal(see Section 4.3). Second, we measure the user engagement rate with a brand CBER 2 as an average of the total number of user engagement rate for all the posts of a specific brand CP ER (1)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "CBER = ( All P osts CP ER)/N P (2)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "Where N P is the total number of posts per Brand.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "UGC analysis", "sec_num": "3.3.1" }, { "text": "We perform an analysis related to post characteristics where we have considered:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "\u2022 Content Type (CT) : we considered three classes: Information Info. about product/service, remuneration Renum. where competitions with rewards and offers are proposed, entertainment Enter. any pleasant and hedonic content and other.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "\u2022 Media Type (MT): status, photo, video, link or event. Some media types keep the user more engaged like videos.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "Another fundamental metric called Post Engagement Rate PER is considered. According to Facebook this metric has different ways of measuring. Bellow, the details about two of them :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "P ER = R + C + S f * 100", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "(3)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "P ER = R + C + S Reach * 100", "eq_num": "(4)" } ], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "Where R is the total number of posts' reactions and C is the total number of posts' comments, S is the total number of posts' shares, f is the total number of followers on the day of posting. Although, the second formula gives a more accurate result than the first one, because it uses the Reach metric which is considered as a private data (visible only to the platform and the page owners). And it can't be applicable by simple users. Thus, we will only use the first Formula for this study.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "BGC analysis", "sec_num": "3.3.2" }, { "text": "As reported in Figure 3 , the resulted chosen sample consists of 50 brands/Services pages with 9977 posts. It is worth mentioning that we have extracted posts with their related information that might be essential for our study like the number of comments, shares, and reactions. We have also scraped all posts' comments for 12 pages that belongs to different subcategories and their related user interactions. In total, we obtain around 900K comments.", "cite_spans": [], "ref_spans": [ { "start": 15, "end": 23, "text": "Figure 3", "ref_id": "FIGREF1" } ], "eq_spans": [], "section": "Results and Discussion", "sec_num": "4" }, { "text": "In addition, to fairly assess these statistics, we have compared them to the 50 first well engaged world Brands as a baseline dataset with a similar distribution of Brands. This latter is collected from the Website Ranking the Brands 3 . This Website presents statistics about the world most engaged brands.The data we used from this source was bound to the same period of collection of our Algerian Facebook pages data.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Results and Discussion", "sec_num": "4" }, { "text": "For some other metrics, we have relied on SocialBakers studies and those of Buffer and Buzzsumo 4 which is considered as one of the largest studies, where they analyzed more than 43 millions Facebook posts from the top 20 k brands in the world 5 .", "cite_spans": [ { "start": 244, "end": 245, "text": "5", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Results and Discussion", "sec_num": "4" }, { "text": "In order to interpret the results, we have chosen to separately analyse them on both sides : Users and Brands' Owners.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Results and Discussion", "sec_num": "4" }, { "text": "Concerning the engagement of Brands' Owners through their pages. We have examined three facts: the used media type (MT) in posts, content type (CT) and the post engagement rate metric (PER). Figure 3a reports the distribution of all posts according to their media type (Event, Link, Photo, Status, Video) currently allowed by Facebook. While it is known that the richest media is Video as it describes the product or service better than a photo. We observe that the most used media by our chosen Brands is the \"Photo\" with more than 85% while only 8.5% of posts deploy videos. In addition, video posts are less used on the Algerian Brand/Service posts counter to the world baseline one with 46%. Furthermore, Event Type is the less used. Figure 2 reports the average post Engagement Rates in term of MT for both Algerian's and the worlds' Brands/Services. According to these comparative results, we notice that Photos, which are the most used MT in Algerian brand/service posts, gives them the highest ER compared to others. We notice a considerable ER for videos and status. Even though they are less used.", "cite_spans": [], "ref_spans": [ { "start": 191, "end": 200, "text": "Figure 3a", "ref_id": "FIGREF1" }, { "start": 738, "end": 746, "text": "Figure 2", "ref_id": null } ], "eq_spans": [], "section": "Brand's Owner Engagement", "sec_num": "4.1" }, { "text": "Comparing the curve of Algerian ER in we observe that link MT gives a considerable ER for the Algerian sample, while this is not the case for the world baseline sample, counter to the event MT which gives high ER for worlds' Brand/Service posts and low ER for the Algerian ones. Figure 3b reports the distribution of all posts according to the used Content type. These results show that most of Algerian Brand/Service posts have Entertainment Content Type, and just 8% of them have Remuneration Content Type. While we notice that information posts are the most published posts in the World baseline brands/services (60%) while they are less considered in the Algerian ones.", "cite_spans": [], "ref_spans": [ { "start": 279, "end": 288, "text": "Figure 3b", "ref_id": "FIGREF1" } ], "eq_spans": [], "section": "Brand's Owner Engagement", "sec_num": "4.1" }, { "text": "Comparing the curves of the worlds' ones in Figure 4b and the Algerian's ones in Figure 4a in terms of post ER according to the CT, we notice that they have the same magnitude. In fact, the remuneration is the most attractive CT followed by Information Content Type for both Algerians' and worlds' posts samples.", "cite_spans": [], "ref_spans": [ { "start": 44, "end": 53, "text": "Figure 4b", "ref_id": "FIGREF2" }, { "start": 81, "end": 90, "text": "Figure 4a", "ref_id": "FIGREF2" } ], "eq_spans": [], "section": "Brand's Owner Engagement", "sec_num": "4.1" }, { "text": "By comparing the results reported in Figure 3a with those in Figure 4a , we notice that the most used CT is \"Entertainment\". However, the \"Remuneration\" and the \"Information\" ones bring higher Post ER than those of \"Entertainment\" in Algerian brand/service posts.", "cite_spans": [], "ref_spans": [ { "start": 37, "end": 43, "text": "Figure", "ref_id": null }, { "start": 61, "end": 70, "text": "Figure 4a", "ref_id": "FIGREF2" } ], "eq_spans": [], "section": "Brand's Owner Engagement", "sec_num": "4.1" }, { "text": "In summary, the engagement of Algerian brands' owners is quite significant.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Brand's Owner Engagement", "sec_num": "4.1" }, { "text": "For the user side involvement in social marketing, we have analyzed their interest through the quantitative and qualitative measure of interactions. In fact, a users' (potential consumer) comments can provide a better feedback and more information. Figure 5a and Figure 5b report the distribution of users' interactions by type of deployed Social Signal and the distribution of emotional reactions, respectively.", "cite_spans": [], "ref_spans": [ { "start": 249, "end": 258, "text": "Figure 5a", "ref_id": "FIGREF3" }, { "start": 263, "end": 272, "text": "Figure 5b", "ref_id": "FIGREF3" } ], "eq_spans": [], "section": "User' Engagement", "sec_num": "4.2" }, { "text": "We observe that users mainly use Reactions (more than 73%), while comments are used with a distribution of 23%. However, users are less active on sharing action. Even though sharing is considered as a deep level of engagement (Aldous et al., 2019) .", "cite_spans": [ { "start": 226, "end": 247, "text": "(Aldous et al., 2019)", "ref_id": "BIBREF1" } ], "ref_spans": [], "eq_spans": [], "section": "User' Engagement", "sec_num": "4.2" }, { "text": "A more fine analysis of emotional reactions, shows (Figure 5b ) that Algerian users are less used to emotional reactions. This latter can be explained by the fact that the emotional reactions has been just introduced by Facebook ", "cite_spans": [], "ref_spans": [ { "start": 51, "end": 61, "text": "(Figure 5b", "ref_id": "FIGREF3" } ], "eq_spans": [], "section": "User' Engagement", "sec_num": "4.2" }, { "text": "In addition to the previous statistical analysis of the collected data. We performed some textual data analytics using Topic Modeling. The aim of Topic Modeling here is to discover the notable topics discussed in comments and quantify to what extent users are engaged with a brand/service content. Differently put, Topic Modeling investigates the comments' relevance by checking similarities between the top words of the topic describing a brand/service and the brand/service related comments. t In the current work, we used Latent Dirichlet Allocation (LDA) method (Blei et al., 2003) that exhibits smooth scalability applied on large textual corpora.", "cite_spans": [ { "start": 566, "end": 585, "text": "(Blei et al., 2003)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "User's content-based engagement", "sec_num": "4.3" }, { "text": "Topic modeling here serves as an aggregation tool to discover the latent discussed topics in users' comments.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "User's content-based engagement", "sec_num": "4.3" }, { "text": "By examining the resulting topics presented in Table 4 , we can clearly notice that they represent most of the Brand/service categories covered by this study.", "cite_spans": [], "ref_spans": [ { "start": 47, "end": 54, "text": "Table 4", "ref_id": "TABREF3" } ], "eq_spans": [], "section": "User's content-based engagement", "sec_num": "4.3" }, { "text": "For instance, most of the words in Topic 0 are related to the Algerian Airlines company (i.e, most of the aggregated words are from AirAlgerie Facebook page. For example, The words:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "User's content-based engagement", "sec_num": "4.3" }, { "text": ", alger, paris, billet, bonne, prix, vol, air, are respectively in English: Allah (the God), Algiers (the capital of Algeria), Paris (the capital of France), Ticket (the flight ticket), good or nice (we assume that it is a typo in writing \"bon\" from \"bon voyage\", Nice trip), Price, Flight and Air from the company name AirAlgerie.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "User's content-based engagement", "sec_num": "4.3" }, { "text": "Another example, is Topic 4 which is about the two biggest telecommunication companies in Algeria: Ooredoo and Mobilis. Most of the words indicating that users' comments are about these companies services. Some of these words are: ,ooredoo, prix , max , win, da. The English translation of the previous words, respectively is: Dzd ( the Algerian currency), valid(most probably about the phone credit), done(a com- mon word used in the Algerian content indicating that someone has seen the post), Mobilis (the first company name), giga (or gigaoctet: a unit for computer memory), Ooredoo (the second company name), Price, Max and Win(plans names offered by the company), Dzd(the abbreviation of tha algerian curruncy in french; Dinar Algerien ). We can also highlight the repetition of the word \" \" in most of the topics' word set. A possible interpretation could be that it is a very common for Algerian users to overuse the expression \" \" English: God willing with all its variations, in their daily life, thus in their online comments .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "User's content-based engagement", "sec_num": "4.3" }, { "text": "Concerning the used languages, Figure 6 illustrates their distribution in users' comments. It shows that French is the most used language, followed by MSA Arabic with 35.7% then Algerian dialect with 19.4%. The category \"Other\" includes Tamazight, Espagnol, Korean and others.", "cite_spans": [], "ref_spans": [ { "start": 31, "end": 39, "text": "Figure 6", "ref_id": "FIGREF4" } ], "eq_spans": [], "section": "Used Languages and Scripts", "sec_num": "4.4" }, { "text": "While for the distribution of used Scripts, we have classified comments according to Arabic scripts, Latin scripts, mixed Arabic, Latin scripts and other scripting sets like emoticons or numbers. We observe that 65% of textual comments use Latin characters while 25% of them use Arabic ones.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Used Languages and Scripts", "sec_num": "4.4" }, { "text": "Moreover, we have analyzed the usage of Emojis where we have reported that just 7% of Algerian customers' comments are using emoticons while 93% of them are textual.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Used Languages and Scripts", "sec_num": "4.4" }, { "text": "These findings could help brands in term of marketing get closer to their customers through understanding their language. In addition, these findings help any Natural Language Processing research problem to leverage such linguistic features. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Used Languages and Scripts", "sec_num": "4.4" }, { "text": "We proposed an analytical study based on statistical and textual analysis of User/Brand Generated Content on social media. We investigated the level of Users/Brands engagement by two means: using the common User engagement formulas in the literature for reactions case. And we suggested a content-based user engagement approach based on LDA Topic Modeling method. Our finding highlight the quantitative and qualitative significance of the existing social signals in the Algerian productivity context. Which can efficiently help", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion", "sec_num": "5" }, { "text": "www.socialbakers.com : social media analytics platform.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "http://gs.statcounter.com/#social_ media-DZ-monthly-201601-201701-bar", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "https://www.rankingthebrands.com\" 4 https://buzzsumo.com/ 5 https://buffer.com/7", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null } ], "back_matter": [ { "text": "Brands' owners to improve their productivity and online marketing strategy. In the future, we intend to normalise the whole content of the data set by, automatically detecting the language and translate it to Modern Standard Arabic. We will also investigate the effectiveness of Topic Modeling on assessing users' engagement.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "acknowledgement", "sec_num": null } ], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "Engagement and brand loyalty through social capital in social media", "authors": [ { "first": "Mohammad", "middle": [], "last": "Abuljadail", "suffix": "" }, { "first": "Louisa", "middle": [], "last": "Ha", "suffix": "" } ], "year": 2019, "venue": "International Journal of Internet Marketing and Advertising", "volume": "13", "issue": "3", "pages": "197--217", "other_ids": { "DOI": [ "10.1504/IJIMA.2019.102557" ] }, "num": null, "urls": [], "raw_text": "Mohammad Abuljadail and Louisa Ha. 2019. Engage- ment and brand loyalty through social capital in so- cial media. International Journal of Internet Mar- keting and Advertising, 13(3):197-217.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "View, like, comment, post: Analyzing user engagement by topic at 4 levels across 5 social media platforms for 53 news organizations", "authors": [ { "first": "Kholoud", "middle": [], "last": "Khalil Aldous", "suffix": "" }, { "first": "Jisun", "middle": [], "last": "An", "suffix": "" }, { "first": "Bernard", "middle": [ "J" ], "last": "Jansen", "suffix": "" } ], "year": 2019, "venue": "Proceedings of the International AAAI Conference on Web and Social Media", "volume": "13", "issue": "", "pages": "47--57", "other_ids": {}, "num": null, "urls": [], "raw_text": "Kholoud Khalil Aldous, Jisun An, and Bernard J. Jansen. 2019. View, like, comment, post: Analyz- ing user engagement by topic at 4 levels across 5 social media platforms for 53 news organizations. Proceedings of the International AAAI Conference on Web and Social Media, 13(01):47-57.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "Diversity in Online Advertising: A Case Study of 69 Brands on Social Media", "authors": [ { "first": "Jisun", "middle": [], "last": "An", "suffix": "" }, { "first": "Ingmar", "middle": [], "last": "Weber", "suffix": "" } ], "year": 2018, "venue": "", "volume": "", "issue": "", "pages": "38--53", "other_ids": { "DOI": [ "10.1007/978-3-030-01129-1_3" ] }, "num": null, "urls": [], "raw_text": "Jisun An and Ingmar Weber. 2018. Diversity in Online Advertising: A Case Study of 69 Brands on Social Media, pages 38-53.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Customer engagement and advertising effectiveness: A moderated mediating analysis", "authors": [ { "first": "Samik", "middle": [], "last": "Anubha", "suffix": "" }, { "first": "", "middle": [], "last": "Shome", "suffix": "" } ], "year": 2021, "venue": "Journal of Internet Commerce", "volume": "0", "issue": "0", "pages": "1--41", "other_ids": { "DOI": [ "10.1080/15332861.2021.1955324" ] }, "num": null, "urls": [], "raw_text": "Anubha and Samik Shome. 2021. Customer engage- ment and advertising effectiveness: A moderated mediating analysis. Journal of Internet Commerce, 0(0):1-41.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Toward a web-based speech corpus for algerian dialectal arabic varieties", "authors": [ { "first": "Soumia", "middle": [], "last": "Bougrine", "suffix": "" }, { "first": "Aicha", "middle": [], "last": "Chorana", "suffix": "" }, { "first": "Abdallah", "middle": [], "last": "Lakhdari", "suffix": "" }, { "first": "Hadda", "middle": [], "last": "Cherroun", "suffix": "" } ], "year": 2017, "venue": "Proceedings of the Third Arabic Natural Language Processing Workshop", "volume": "", "issue": "", "pages": "138--146", "other_ids": {}, "num": null, "urls": [], "raw_text": "Soumia Bougrine, Aicha Chorana, Abdallah Lakhdari, and Hadda Cherroun. 2017. Toward a web-based speech corpus for algerian dialectal arabic varieties. In Proceedings of the Third Arabic Natural Lan- guage Processing Workshop, pages 138-146.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Les m\u00e9dias sociaux: L'\u00e9tude de l'effet m\u00e9diateur de la confiance et l'utilit\u00e9 per\u00e7ue des commentaires dans le contexte alg\u00e9rien", "authors": [ { "first": "Amel", "middle": [], "last": "Graa", "suffix": "" }, { "first": "Soumia", "middle": [], "last": "Abdelhak", "suffix": "" }, { "first": "Hayat", "middle": [], "last": "Baraka", "suffix": "" } ], "year": 2017, "venue": "International Journal of Marketing, Communication and New Media", "volume": "0", "issue": "2", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Amel Graa, Soumia Abdelhak, and Hayat Baraka. 2017. Les m\u00e9dias sociaux: L'\u00e9tude de l'effet m\u00e9- diateur de la confiance et l'utilit\u00e9 per\u00e7ue des com- mentaires dans le contexte alg\u00e9rien. International Journal of Marketing, Communication and New Me- dia, 0(2).", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "Customer engagement factors in facebook brand pages", "authors": [ { "first": "Sudarsan", "middle": [], "last": "Jayasingh", "suffix": "" }, { "first": "Rajagopalan", "middle": [], "last": "Venkatesh", "suffix": "" } ], "year": 2015, "venue": "Asian Social Science", "volume": "11", "issue": "26", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Sudarsan Jayasingh and Rajagopalan Venkatesh. 2015. Customer engagement factors in facebook brand pages. Asian Social Science, 11(26):19.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Measuring user engagement. Synthesis lectures on information concepts, retrieval, and services", "authors": [ { "first": "Mounia", "middle": [], "last": "Lalmas", "suffix": "" }, { "first": "O'", "middle": [], "last": "Heather", "suffix": "" }, { "first": "Elad", "middle": [], "last": "Brien", "suffix": "" }, { "first": "", "middle": [], "last": "Yom-Tov", "suffix": "" } ], "year": 2014, "venue": "", "volume": "6", "issue": "", "pages": "1--132", "other_ids": { "DOI": [ "https://www.morganclaypool.com/doi/pdfplus/10.2200/S00605ED1V01Y201410ICR038?casa_token=IOgVQL2lHNMAAAAA:foabHVvRaXnz6Mz25Cw2UiVxoGk0g3lTpj41rCb8L7whC733eztFDVlBXrPFm32pX0Dz3RgSjEnI4w" ] }, "num": null, "urls": [], "raw_text": "Mounia Lalmas, Heather O'Brien, and Elad Yom-Tov. 2014. Measuring user engagement. Synthesis lec- tures on information concepts, retrieval, and ser- vices, 6(4):1-132.", "links": null }, "BIBREF9": { "ref_id": "b9", "title": "A Proposed Lexicon-Based Sentiment Analysis Approach for the Vernacular Algerian Arabic", "authors": [ { "first": "Mhamed", "middle": [], "last": "Mataoui", "suffix": "" }, { "first": "Omar", "middle": [], "last": "Zelmati", "suffix": "" }, { "first": "Madiha", "middle": [], "last": "Boumechache", "suffix": "" } ], "year": 2016, "venue": "Research in Computing Science", "volume": "110", "issue": "", "pages": "55--70", "other_ids": {}, "num": null, "urls": [], "raw_text": "Mhamed Mataoui, Omar Zelmati, and Madiha Boumechache. 2016. A Proposed Lexicon-Based Sentiment Analysis Approach for the Vernacular Algerian Arabic. Research in Computing Science, 110:55-70.", "links": null }, "BIBREF10": { "ref_id": "b10", "title": "Brand Engagement on Facebook Based on Mobile Phone Operators' Activity Within Four European Countries", "authors": [ { "first": "Artur", "middle": [], "last": "Bernard Olczak", "suffix": "" }, { "first": "Rita", "middle": [ "Karolina" ], "last": "Sobczyk", "suffix": "" } ], "year": 2013, "venue": "Studia Ekonomiczne", "volume": "150", "issue": "", "pages": "97--108", "other_ids": {}, "num": null, "urls": [], "raw_text": "Artur Bernard Olczak and Rita Karolina Sobczyk. 2013. Brand Engagement on Facebook Based on Mobile Phone Operators' Activity Within Four Eu- ropean Countries. Studia Ekonomiczne, 150:97- 108.", "links": null }, "BIBREF11": { "ref_id": "b11", "title": "Social Media Engagement: Content Strategy and Metrics Research Opportunities", "authors": [ { "first": "Marie-", "middle": [], "last": "", "suffix": "" }, { "first": "Catherine", "middle": [], "last": "Perreault", "suffix": "" }, { "first": "Elaine", "middle": [], "last": "Mosconi", "suffix": "" } ], "year": 2018, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": { "DOI": [ "10.24251/HICSS.2018.451" ] }, "num": null, "urls": [], "raw_text": "Marie-Catherine Perreault and Elaine Mosconi. 2018. Social Media Engagement: Content Strategy and Metrics Research Opportunities.", "links": null }, "BIBREF12": { "ref_id": "b12", "title": "Online engagement factors on facebook brand pages", "authors": [ { "first": "Irena", "middle": [], "last": "Pletikosa Cvijikj", "suffix": "" }, { "first": "Florian", "middle": [], "last": "Michahelles", "suffix": "" } ], "year": 2013, "venue": "Social Network Analysis and Mining", "volume": "3", "issue": "4", "pages": "843--861", "other_ids": { "DOI": [ "10.1007/s13278-013-0098-8" ] }, "num": null, "urls": [], "raw_text": "Irena Pletikosa Cvijikj and Florian Michahelles. 2013a. Online engagement factors on facebook brand pages. Social Network Analysis and Mining, 3(4):843-861.", "links": null }, "BIBREF13": { "ref_id": "b13", "title": "Online engagement factors on Facebook brand pages", "authors": [ { "first": "Irena", "middle": [], "last": "Pletikosa Cvijikj", "suffix": "" }, { "first": "Florian", "middle": [], "last": "Michahelles", "suffix": "" } ], "year": 2013, "venue": "Social Network Analysis and Mining", "volume": "3", "issue": "4", "pages": "843--861", "other_ids": { "DOI": [ "10.1007/s13278-013-0098-8" ] }, "num": null, "urls": [], "raw_text": "Irena Pletikosa Cvijikj and Florian Michahelles. 2013b. Online engagement factors on Facebook brand pages. Social Network Analysis and Mining, 3(4):843-861.", "links": null }, "BIBREF14": { "ref_id": "b14", "title": "Araacom: Arabic algerian corpus for opinion mining", "authors": [ { "first": "Hichem", "middle": [], "last": "Rahab", "suffix": "" }, { "first": "Abdelhafid", "middle": [], "last": "Zitouni", "suffix": "" }, { "first": "Mahieddine", "middle": [], "last": "Djoudi", "suffix": "" } ], "year": 2017, "venue": "Proceedings of the International Conference on Computing for Engineering and Sciences, ICCES '17", "volume": "", "issue": "", "pages": "35--39", "other_ids": { "DOI": [ "10.1145/3129186.3129193" ] }, "num": null, "urls": [], "raw_text": "Hichem Rahab, Abdelhafid Zitouni, and Mahieddine Djoudi. 2017. Araacom: Arabic algerian corpus for opinion mining. In Proceedings of the Inter- national Conference on Computing for Engineering and Sciences, ICCES '17, pages 35-39, New York, NY, USA. ACM.", "links": null }, "BIBREF15": { "ref_id": "b15", "title": "Internet social networking", "authors": [ { "first": "Daniel", "middle": [], "last": "Richter", "suffix": "" }, { "first": "Kai", "middle": [], "last": "Riemer", "suffix": "" } ], "year": 2011, "venue": "Business Information Systems Engineering", "volume": "3", "issue": "2", "pages": "89--101", "other_ids": {}, "num": null, "urls": [], "raw_text": "Daniel Richter, Kai Riemer, and Jan vom Brocke. 2011. Internet social networking. Business Information Systems Engineering, 3(2):89-101.", "links": null }, "BIBREF16": { "ref_id": "b16", "title": "Is digital advertising effective under conditions of low attention", "authors": [ { "first": "Irene", "middle": [], "last": "Santoso", "suffix": "" }, { "first": "Malcolm", "middle": [], "last": "Wright", "suffix": "" }, { "first": "Giang", "middle": [], "last": "Trinh", "suffix": "" }, { "first": "Mark", "middle": [], "last": "Avis", "suffix": "" } ], "year": 2020, "venue": "Journal of Marketing Management", "volume": "36", "issue": "", "pages": "1707--1730", "other_ids": { "DOI": [ "10.1080/0267257X.2020.1801801" ] }, "num": null, "urls": [], "raw_text": "Irene Santoso, Malcolm Wright, Giang Trinh, and Mark Avis. 2020. Is digital advertising effective un- der conditions of low attention? Journal of Market- ing Management, 36(17-18):1707-1730.", "links": null }, "BIBREF17": { "ref_id": "b17", "title": "Sentiment analysis of users on social networks: overcoming the challenge of the loose usages of the algerian dialect", "authors": [ { "first": "Assia", "middle": [], "last": "Soumeur", "suffix": "" }, { "first": "Mheni", "middle": [], "last": "Mokdadi", "suffix": "" }, { "first": "Ahmed", "middle": [], "last": "Guessoum", "suffix": "" }, { "first": "Amina", "middle": [], "last": "Daoud", "suffix": "" } ], "year": 2018, "venue": "Procedia computer science", "volume": "142", "issue": "", "pages": "26--37", "other_ids": {}, "num": null, "urls": [], "raw_text": "Assia Soumeur, Mheni Mokdadi, Ahmed Guessoum, and Amina Daoud. 2018. Sentiment analysis of users on social networks: overcoming the challenge of the loose usages of the algerian dialect. Procedia computer science, 142:26-37.", "links": null }, "BIBREF18": { "ref_id": "b18", "title": "Brand communication in social media: A research agenda", "authors": [ { "first": "A", "middle": [ "M" ], "last": "Hilde", "suffix": "" }, { "first": "", "middle": [], "last": "Voorveld", "suffix": "" } ], "year": 2019, "venue": "Journal of Advertising", "volume": "48", "issue": "1", "pages": "14--26", "other_ids": { "DOI": [ "10.1080/00913367.2019.1588808" ] }, "num": null, "urls": [], "raw_text": "Hilde A.M. Voorveld. 2019. Brand communication in social media: A research agenda. Journal of Adver- tising, 48(1):14-26.", "links": null }, "BIBREF19": { "ref_id": "b19", "title": "Understanding user-generated content and customer engagement on facebook business pages", "authors": [ { "first": "Mochen", "middle": [], "last": "Yang", "suffix": "" }, { "first": "Yuqing", "middle": [], "last": "Ren", "suffix": "" }, { "first": "Gediminas", "middle": [], "last": "Adomavicius", "suffix": "" } ], "year": 2019, "venue": "Information Systems Research", "volume": "30", "issue": "3", "pages": "839--855", "other_ids": { "DOI": [ "10.1287/isre.2019.0834" ] }, "num": null, "urls": [], "raw_text": "Mochen Yang, Yuqing Ren, and Gediminas Adomavi- cius. 2019. Understanding user-generated content and customer engagement on facebook business pages. Information Systems Research, 30(3):839- 855.", "links": null }, "BIBREF20": { "ref_id": "b20", "title": "Prediction of the engagement rate on algerian dialect facebook pages. Recent Advances in NLP: The Case of Arabic Language", "authors": [ { "first": "Chayma", "middle": [], "last": "Zatout", "suffix": "" }, { "first": "Ahmed", "middle": [], "last": "Guessoum", "suffix": "" }, { "first": "Chemseddine", "middle": [], "last": "Neche", "suffix": "" }, { "first": "Amina", "middle": [], "last": "Daoud", "suffix": "" } ], "year": 2019, "venue": "", "volume": "874", "issue": "", "pages": "", "other_ids": { "DOI": [ "https://link.springer.com/chapter/10.1007/978-3-030-34614-0_9" ] }, "num": null, "urls": [], "raw_text": "Chayma Zatout, Ahmed Guessoum, Chemseddine Neche, and Amina Daoud. 2019. Prediction of the engagement rate on algerian dialect facebook pages. Recent Advances in NLP: The Case of Arabic Lan- guage, 874:163.", "links": null } }, "ref_entries": { "FIGREF0": { "type_str": "figure", "uris": null, "num": null, "text": "Figure 1: Methodology Overview" }, "FIGREF1": { "type_str": "figure", "uris": null, "num": null, "text": "Figure 2a and the world baseline one in Figure 2b, (a) Average of ER received by Algerian Brands/Services (b) Average of ER received by World Brands/Services Figure 2: Distribution of Posts & ER by MT (Algeria vs. World) Distribution of Posts by (a) Media Type, (b) Content Type." }, "FIGREF2": { "type_str": "figure", "uris": null, "num": null, "text": "Distribution of Posts & ER by CT (Algeria vs. World) at the time of collection of our dataset." }, "FIGREF3": { "type_str": "figure", "uris": null, "num": null, "text": "Distribution by type of (a) Social Signal (b) Emotional Reactions" }, "FIGREF4": { "type_str": "figure", "uris": null, "num": null, "text": "Distribution of used languages in User comments" }, "TABREF0": { "html": null, "text": "Details on some User/Brand Engagement studies. Media type, posting day and time (Olczak and Sobczyk, 2013) 10 pages belongs to 4 mobile brands Facebook Number of likes, number of shares and posting time.", "type_str": "table", "num": null, "content": "
WorkDatasetPlatformMetrics and Factors
(Pletikosa Cvijikj and
Michahelles, 2013a) Content type, (Jayasingh and Venkatesh, 100 Brand pages Facebook 10 169 Posts of
2015)134 Brand pagesFacebookNumber of fans, Customer interaction and Posts type
12K posts of
business pages of
500 companies in 6Number of likes and posts' linguistic features, poster
(Yang et al., 2019)industriesFacebookcharacteristics, post context heterogeneity.
Facebook,
Instagram,
3 M social postsTwitter,
from 53 newsYouTube, and
(Aldous et al., 2019)organizationsRedditshares, external posting, Topic variations
\u2022 How are Algerian Brand owners exploit-
ing Social Media?
" }, "TABREF1": { "html": null, "text": "Corpora for Algerian Social Data.", "type_str": "table", "num": null, "content": "
CorpusPurposeCorpus DetailsAvailable
Algerian
Lexicon (Mataoui et al.,206 posts, 7698 comments, Manually
2016)Sentiment Analysiscollected and annotatedNo
ARAACOM (Rahab
et al., 2017)Opinion MiningComments on Algerian newspaperNo
20 Algerian brand pages, 25475
(Soumeur et al., 2018)Sentiment Analysisannotated comments.No
and Sobczyk, 2013; Yang et al., 2019; Aldous
et al., 2019). Table 1 gives some details on the
used metrics and factors. The salient remark
is that most used metrics are based on quanti-
tative measurements, namely, the number of
reactions and posting times. For news organ-
isations, Aldous et al. (Aldous et al., 2019)
defined a more efficient engagement metric
based on the user behavior leading to exter-
nal posting (Spreading content through public
sharing to other public networks or platforms).
This is performed by means of studying topic
variation.
" }, "TABREF2": { "html": null, "text": "Details on Chosen Brands and Services. We define the main keywords that can help automatically search targeted lists. When such lists are established, a first filtering is performed to keep only the potential suitable data. It helps to enlarge our Brand-list by Brands that are well visible via Social networks (i.e. well ranked) but not considered by experts as a powerful Brand/Service. (ii)Downloading Data: in this step, we use customized scripts, and Facebook Graph API to scrape the data.", "type_str": "table", "num": null, "content": "
CategorySubcategory#Illustration#Post#Comment
Condor Electronics, ENIEM, Cobra
BrandAppliance5Electronics, ENIE, StartLight1 24721 786
Beauty/Hygiene4Awane,Bimbies,finessecepro,Venus577
CAFE-Boukhari, Aroma-Caf\u00e9, Rouiba-Jus,
Beverage6Vita-Jus, Cevital-boissons, Ngaous1 10615 305
Soummam, FALAIT-Tartino,
Dairy3Berber-fromage423
LG Algerie, Oppo Algerie,
Electronics/Phone4HuaweimobileDZ, SonymobileDZ937
Benamor, Safina, Sim, CevitalCulinaire,
Food6Jumbo, Bimo1451
Furniture2Dz-meuble, Sotrabois menuiserie d'art19931 398
Nassah, El-Bahdjadetergents, Aigle, Force
Household Goods4Xpress34726 576
Imetal-SIDER EL-ADJAR, SNVI,
Industrial4TEXALG ex. Sonitex, ENAP81256
Industrial/Auto2Renault'DZ, Dacia'DZ
ServicesAccommodation3El-Djazair, ElAurassi, El Biar hotel5434
Telecommunication3Djezzy'DZ, Mobilis, Ooredoo'DZ1 960665 284
Transportation/Airlines 2Air Algerie, Tassili Airlines25735 274
Web Service1Ouedkniss.com56545 539
Total14509 977906 705
1. Inventory of Potential Algerianothers. For example, in the time span of
Brands/Products/Services :this study Algerian users are less inter-
First, Brands/Products/Services we have the most representative of Algerian identified that are productivity. This is mainly done using direct expert advice and some social media analytic platforms such as Social-Bakers 1 . This step leads to a preliminary list of Brands and services.ested in Instagram or Snapchat compared to Middle Est and Gulf communities. In fact, they commonly use Facebook and YouTube 2 . These statistics show that from the period between January and Novem-ber 2017(the period of our dataset collec-tion), Facebook represents the most used social media platform with 75.94% fol-lowed by Youtube and Twitter with only
2. Inventory of Potential social Media11.37% and 8.28% respectively.
sources:3. Extraction Process
we have identified the common social me-In order to avoid collecting useless data.
dia platforms used by communities in con-This step is achieved in two stages:
cerns. Indeed, depending on their culture(i)Providing Lists:
and preferences, some communities show
preferences of some social media over
" }, "TABREF3": { "html": null, "text": "Example of generated topwords of a topic using LDA model.", "type_str": "table", "num": null, "content": "
Topic
NumberTop wordsTop words(English)Description
air,, alger, paris, billet, bonne, algerie,air,Allah(God), Algiers, Paris, ticket, goodthis represents the airlines
bien, prix, share, site, mohamed, saha, da,Algeria, price, share, site, Mohamed, okay,Facebook page
Topic 0vol,Dzd, flight, days
hada, aroma , chaba , Mohamed ,, top ,This,aroma,nice, Mo-hamed,coffee,top,price,lottery,fantastic,aromaThis is the Aroma coffee
Topic 1prix ,,,,God willing, doneFacebook page
Thanks, LG, participate, luck, product, force,
merci, lg, participe, chance, produits, force,price, Aigle, good, top , express, wash,The cleaning stuff topic for
prix, aigle, bien, top, express, lave,xpress, congratulation, LG, Algeria, Allah,AigleGroupt, and Force
Topic 2xpress,thanksExpress Brands
ooredoo, prix , max , win, da, go ,Ooredoo, price, max , win, Dzd, GO, valid, an hour, Dzd, octet, , the answer, fee, month,This topic is about telecommunication companies
Topic 3Giga, done,la win, MobilisOoredoo and Mobilis
" } } } }