A Sentiment Analysis Visualization System for the Property Industry

The usage of social media platforms such as Facebook and Twitter, either by the public or by organizations, has been rapidly increasing. The decision-makers in the organizations use social media to engage with their customers since public users tend to express their opinions about certain products and services through this popular mechanism. Hence, this valuable data can be useful for marketing and business decisions. However, the main obstacle is obtaining meaningful information from these platforms due to the unstructured data they present. Sentiment analysis is seen as the best tool to analyze insights or opinions in this huge amount of data. In this article, we extract data on public opinion about property in order to understand the reason behind the imbalances of supply and demand currently faced by the property industry in Malaysia. In addition, we visualized the sentiment results in the form of a dashboard so that it may help property players to understand the public sentiments toward their housing or construction projects. 

Housing; Property; Sentiment analysis; Social media; Visualization

The property scenario in Malaysia is currently facing a crucial supply and demand imbalance. As reported by the Central Bank of Malaysia (BNM), the supply and demand imbalances in the property market have increased since 2015, with unsold residential properties already at their highest in 10 years. According to the leading property consultant, Knight Frank Malaysia, Malaysia’s property market is expected to be sluggish this year (NST Business, 2018). This problem is due to the oversupply situation and the imbalances in supply and demand for the properties. The causes of the supply and demand mismatch crucially need to be addressed to ensure that the property market can completely reach its target.

Therefore, the BNM proposed that the government should gather information about characteristics and preferences of the public in order to effectively meet the demand of households. The lack of that information has contributed to a large number of overhang properties, including the affordable housing projects in Johor, Selangor, and Kedah (Kay, 2018). In addition, Datuk Seri FD Iskandar Mohamed Mansor, the president of the Real Estate and Housing Developers’ Association (REHDA), also suggested that the property and real estate players should have a one-stop data center to assist decision-makers in the property industry to better bridge the gap between supply and demand and to predict future housing trends (Ling et al., 2017). It is crucial to have an integrated database on property supply and demand that can provide insights on the needs and preferences of households, their links to demographic shifts, and property gaps across Malaysia (Mustafa et al., 2017).

Currently, several studies involving traditional methods of research, such as surveys and questionnaires that targeted only certain groups of people have been conducted to gather public preferences on the property sector, including affordable housing projects (Jamaluddin et al., 2016). However, the traditional research methods are limited to a particular set of questions that are sometimes forced onto people who might not give candid answers. To address this gap, we proposed to gather information from social media platforms since the amount of data that can be captured through Twitter and Facebook is massive. Moreover, the information shared in social media is considered to be honest feedback from the public since they posted their opinions without being asked for it.

In order to analyze such a huge amount of unstructured data in social media, we used Sentiment Analysis (SA) as a tool to compute and analyze public opinions or sentiments written in the form of text. SA is a field of research in Natural Language Processing (NLP) that aims to automatically determine the attitude of a speaker or writer based on the subjective information shared on the Web (Pang & Lee, 2008; Liu, 2012). The importance of this field has been proven by the extensive number of methods and approaches that have been proposed in research as well as by the interest of organizations and companies that it has raised in recent years. Previous studies have reported that SA has been applied on wide variety of topics and issues such as online products reviews (e.g., mobile phones; Di Fabbrizio et al., 2013), hotel reviews (Kasper & Vela, 2011), political and financial analysis (Soelistio & Surendra, 2015; Chiong et al., 2018), housing (Mahadzir et al., 2016) and the prediction of real-word events (Rifai et al., 2015). Past research on the data visualization of SA in various domains, including political sentiments during 2012 United States presidential election (Wang et al., 2012), real-time monitoring and analysis of football (Saavedra, 2016), and customer reviews on products and services (Al Kubaizi et al., 2018; Chen & Zheng, 2018), have been presented.

SA is commonly divided into two main techniques, which are machine learning and lexicon based. The machine learning technique attempts to train a sentiment classifier based on the occurrence frequencies of the various words in the datasets (Feldman, 2013; Santosh & Vardhan, 2015). There are several well-known machine learning methods that have been applied such as Maximum Entropy, Support Vector Machine (SVM), and Naïve Bayes. Based on previous study, it has been demonstrated that the Naïve Bayes method leads to better performance and accurate classification (Kunal et al., 2018). Meanwhile, the sentiment lexicon requires sentiment dictionaries consisting of sentiment words and their polarity to classify words. For example, the polarity for “best” and “worst” are positive and negative respectively. Various sentiment lexicons have been constructed either manually or semi-automatically such as SentiWordNet (Medagoda et al., 2015) and SenticNet (Cambria et al., 2014).

A huge amount of previous research has been done in mining the sentiments written in the English language. Despite the fact that SA research in English is rather mature, SA studies in other languages such as Malay have just set sail (Al-Moslmi et al., 2017).

In this paper, a case study method is presented that uses Twitter data to analyze public sentiments toward the Perumahan Rakyat 1Malaysia (PR1MA) project. We proposed the implementation of SA toward the property industry using the machine learning method. We used Naïve Bayes as a classifier due to its accurateness and effectiveness as demonstrated in previous Malay SA research (Alshalabi et al., 2013). In order to make the analysis results readable and understandable by the property players, we visualized the results in the form of a dashboard. The dashboard is able to help the property players in understanding public preferences and refining their marketing strategies to enable the industry to better bridge the gap between supply and demand in this sector. 

This SA visualization system is meant to assist the property players in understanding the public views about their housing or construction projects. Hence, decision-makers could make subtle and informed decisions with a better understanding of outside information.

In this paper, we proposed the implementation of SA toward the property industry. A case study involving Twitter data in analyzing public sentiments toward the PR1MA project has been presented. We implemented a machine learning algorithm and used Naïve Bayes to carry out the sentiment classification process. To make the analysis results readable and understandable by the property players, we visualized the results in the form of a dashboard. Our dashboard system consists of three main components which are the overall and feature-based SA, Twitter data monitoring, and reporting.

However, there is a room for improvement. In the future, we are planning to apply various algorithms or to use the sentiment lexicon to perform the classification instead of machine learning alone. Our goal is to find the best classification technique for Malay and English text. We also will extend our work by extracting data from other social media platforms such as Facebook and online forums.

This research is supported by the University Grant (S/O Code: 13879), University Utara Malaysia, 2018.

Al Kubaizi, R., Al-Otaibi, S., Al Washigry, B., Al Suhaim, E., Al Sughayer, J., Al Jumaiah, R., 2018. Mining Expertise using Social Media Analytics. In: 1^st International Conference on Computer Applications & Information Security (ICCAIS), IEEE, pp. 1–5

Al-Moslmi, T., Omar, N., Albared, M., Alshabi, A., 2017. Enhanced Malay Sentiment Analysis with an Ensemble Classification Machine Learning Approach. Journal of Engineering and Applied Sciences, Volume 12(20), pp. 5226–5232

Alshalabi, H., Tiun, S., Omar, N., Albared, M., 2013. Experiments on the Use of Feature Selection and Machine Learning Methods in Automatic Malay Text Categorization. Procedia Technology, Volume 11, pp. 748–754

Cambria, E., Olsher, D., Rajagopal, D., 2014. SenticNet 3: A Common and Common-sense Knowledge Base for Cognition-driven Sentiment Analysis. In: AAAI’14 Proceedings of the  Twenty-eighth AAAI Conference on Artificial Intelligence, Québec City, Québec, Canada, pp. 1515–1521

Chen, C.I.P., Zheng, J., 2018. Improved Big Data Analytics Solution using Deep Learning Model and Real-time Sentiment Data Analysis Approach. In: International Conference on Brain Inspired Cognitive Systems, Springer, Cham, pp. 579–588

Chiong, R., Fan, Z., Hu, Z., Adam, M.T.P., Lutz, B., Neumann, D., 2018. A Sentiment Analysis-based Machine Learning Approach for Financial Market Prediction Via News Disclosures. In: Proceedings of the Genetic and Evolutionary Computation Conference Companion, Kyoto, Japan, pp. 278–279

Di Fabbrizio, G., Stent, A.J., Gaizauskas, R., 2013. Summarizing Opinion-related Information for Mobile Devices. In: Mobile Speech and Advanced Natural Language Solutions, Springer, New York, pp. 289–317

Feldman, R., 2013. Techniques and Applications for Sentiment Analysis. Communications of the ACM, Volume 56(4), pp. 82–89

Jamaluddin, N.B., Abdullah, Y.A., Hamdan, H., 2016. Encapsulating the Delivery of Affordable Housing: An Overview of Malaysian Practice. In: MATEC Web of Conferences, Volume 66

Kasper, W., Vela, M., 2011. Sentiment Analysis for Hotel Reviews. In: Computational Linguistics-Applications Conference, Jachranka, Poland, pp. 45–52

Kay, L.K., 2018. Rehda: Malaysia's Property Industry Must Tap into Big Data. Available Online at https://www.edgeprop.my/content/1289419/rehda-malaysia’s-property-industry-must-tap-big-data, Accessed on 10 September 2018

Kunal, S., Saha, A., Varma, A., Tiwari, V., 2018. Textual Dissection of Live Twitter Reviews using Naive Bayes. Procedia Computer Science, Volume 132, pp. 307–313

Ling, C.S., Almeida, S.J., Wei, H.S., 2017. Affordable Housing: Challenges and the Way Forward. Bulletin. Bank Negara Malaysia. Available Online at http://www.bnm.gov.my/files/publication/qb/2017/Q4/p3ba1.pdf, Accessed on 15 September 2018

Liu, B., 2012. Sentiment Analysis and Opinion Mining. Synthesis Lectures on Human Language Technologies, Volume 5(1), pp. 1–167

Mahadzir, N. H., Omar, M. F., Nawi, M. N. M. 2016. Towards Sentiment Analysis Application in Housing Projects. Journal of Telecommunication, Electronic and Computer Engineering, Volume 8(8), pp. 145-148

Medagoda, N., Shanmuganathan, S., Whalley, J., 2015. Sentiment Lexicon Construction using SentiWordNet 3.0. In: 11^th International Conference on Natural Computation (ICNC), Zhangjiajie, China, pp. 802–807

Mustafa, A., Adnan, N., Nawayai, S.S.M., 2017. The Influence of Product Quality and Service Quality on House Buyer's Satisfaction in Prima Home. Pertanika Journal of Social Science and Humanities, Volume 25(4), pp. 1841–1851

NST Business, 2018. Local Property Market Continues to Self-correct: Knight Frank Malaysia. Available Online at https://www.nst.com.my/business/2018/01/329939/local-property-market-continues-self-correct-knight-frank-malaysia, Accessed on July 20, 2018

Pang, B., Lee, L., 2008. Opinion Mining and Sentiment Analysis. Foundations and Trends in Information Retrieval, Volume 2(1-2), pp. 1–135

Rifai, A.I., Hadiwardoyo, S.P., Correia, A.G., Pereira, P., Cortez, P., 2015. The Data Mining Applied for the Prediction of Highway Roughness due to Overloaded Trucks. International Journal of Technology, Volume 6(5), pp. 751–761

Saavedra, A.P., 2016. Development of a Dashboard for Sentiment Analysis of Football in Twitter based on Web Components and D3. js. Master’s Thesis, ETSI Telecommunication, Universidad Politécnica de Madrid, Spain

Santosh, D.T., Vardhan, B.V., 2015. Obtaining Feature- and Sentiment-based Linked Instance RDF Data from Unstructured Reviews using Ontology-based Machine Learning. International Journal of Technology, Volume 6(2), pp. 198–206

Soelistio, Y.E., Surendra, M.R.S., 2015. Simple Text Mining for Sentiment Analysis of Political Figure using Naive Bayes Classifier Method. In: The Proceedings of The 7^th ICTS, Bali, Indonesia

Wang, H., Can, D., Kazemzadeh, A., Bar, F., Narayanan, S., 2012. A System for Real-time Twitter Sentiment Analysis of 2012 US Presidential Election Cycle. In: Proceedings of the 50^th Annual Meeting of the Association for Computational Linguistics, Jeju, Republic of Korea, pp. 115–120

Yu, L.C., Lee, C.W., Pan, H.I., Chou, C.Y., Chao, P.Y., Chen, Z.H., Tseng, S.F., Chan, C.L., Lai, K.R., 2018. Improving Early Prediction of Academic Failure using Sentiment Analysis on Self?evaluated Comments. Journal of Computer Assisted Learning, Volume 34(4), pp. 358–365