https://indify.co/widgets/live/quotes/hwP6ajJchIIalfbKo0LZ
<aside> <img src="/icons/chat-user_gray.svg" alt="/icons/chat-user_gray.svg" width="40px" />
</aside>
Abstract— Social network is generally regarded as the leading cause of new interaction online together with the freedom of speech incorporated. At the same time, the development of this platform brings up the opportunity to enhance many domains' demands in understanding user behavior rapidly. This work provides insights into the application of crowdsourcing towards different places to respond on urban centralization and decentralization issues. The approach is primarily based on the Twitter data shouted by the user, then being identified in three different text analytical measurement groups. Moreover, this study will visualize the resulting data back to the intended post location as the decision-making indicator created by the users. Taken together, these findings demonstrate the possibility to collect the user's data online without passing the boundary of individual privacy. Nevertheless, the outcome provided can contribute to the prediction of smart city policy decisions in the society's initial stage.
Keywords—crowdsourcing, centralization, decentralization, smart city, twitter, decision-making
The increased growth of social media interaction in society creates an opportunity for people to improve their living quality. Started by sharing activities online, recently, social network has become platform to accommodate user on freely talk and share their thought in the community. Some of the applications of social media accommodate the user many interactions a part of the daily activities conducted in common. Therefore, taking benefit of this fact, many stakeholders are willing to utilize the data taken from the internet as a foundation strategy of their business. The aim of this work is to broaden our knowledge of the objectivity of data provided by the users in particular condition. The investigation of the Twitter user was carried out to determine the centralization and decentralization of data gathered on public. Moreover, the crowdsourcing implemented promotes the data-driven evidence of making the early decision in the application of smart city. The outcomes of this trial will have serious implications for leveraging the usage of data that not only based on the attitudinal report, but also objective behavior of the citizen perceptions across the internet.
The smart city has been emerging and is triggering to improve the convenience of urban management in many assessments. Along with the growth of the trend, the concept merges with the technology application to ease the process of managing the data in the city appropriately, including an enabling urban computing to improve the innovation and development of the city [1]. However, with a lot of data produced in the city, it is a great opportunity to properly enable information decision making to support the sustainability and guidance of the smart city and its sub-application [2].
Obtaining the need of contribution to develop the future smart city that covers all the citizen demands as the insight. Crowdsourcing is an essential concept to establish new relationships and interactive services with the aim of optimizing probable technology and innovation towards people and social inclusion for the urban innovative plan [3]. Nevertheless, the benefit of utilizing social media recently has eased the process of collecting the data digitally. The privacy that has been becoming a concern is lessening the border due to the willingness of people to share their thoughts and activities online. Wang et al. [4] implemented the BE-specific term construction and expansion method to address specific land use problems in different urban contexts, the result claimed the use of social networks as a platform to examine the thoughts and perceptions of citizens in the city. This research also finds that Twitter is a useful source of information to gain the environmental perception and formalizing decision-making knowledge for crowdsourcing. With the same mechanism Zhang et al. [5] examined 28 researches that applied social networks (Twitter) as part of decision-making scenarios. The research emphasizes problem formulation, solution, feature, and information transformation methods to overcome related issues regarding product sales prediction, stock selection, crime prevention, epidemic tracking, and traffic monitoring. The comprehensive analysis deployed in this research presents the usage of social media as a consideration platform for decision maker to encompass multidomain and application in the society. Furthermore, the use of the crowdsourcing scenario is crucial for the fundamental process in a smart city to collect data and develop the service layers [6].
The approach city organizes the various systems and functions refer to the context of centralization and decentralization system applied. It is becoming a tendency for the high-income countries to distribute the decision making into more localize and regional-based [7]. In the city management system, centralization and decentralization cannot be independently applied without any consideration involved. A study conducted in Afghanistan by Mushkani and Ono [8] approved the use of both contexts in the development of urban management, which can be envisioned as centralization in relation to the domestic power guides of biopower, whereas decentralization contributes to the supranational power guides for urban development. Moreover, in the field of technology application, the concept of smart city that involves a blockchain technology also presents the existing decentralization context in enhancing the reliability and efficiency of the secured system used [9]. With the opposite decision, Castillo-Manzano et al. [7] mentioned that Spain enables to demonstrate the application of public centralization applied in regional traffic law. The research showed the low traffic accident and mortality figures under national policing in compared to the local police governance. Under this policy, Spain government reflects to the usage of resources that can be allocated to the costly instrument spreading and the traffic management system particularly in the special tourist season. When it is being compared, Cheng et al. [10] verified the utilization of centralization and decentralization towards the electric vehicle charging port with taking the case study apart of California electric grid in 2030. Above all of the benefits of decentralization to grant the efficiency for the users, this research provided an evidenced that the CO2 emissions reduces in similar levels to the centralized smart charging decision in certain conditions. In summary each context deserves a good proportion dependable on the case study implemented; nonetheless, our research is looking for more objective decision in collaborating the initial viewpoint of the implication issue toward the citizen as user-based decision making.
The social media in the recent days play an important role as a platform that provides insights, needs, and concerns from citizens in the city. In addition, social media also accommodate more informed data decisions and community-based issues as a part of reflective action of the population in the city. As one of existed social medias, twitter provides many benefits to access the user's data for many purposes. According to Arku et al. [11] the capability of twitter as a source of data brings a strong, consistent connection between tweeted-words related to anticipation, trust, and joy and the emotion being expressed. Using the NRC Emotion Lexicon, the research evaluated the effects of marketing campaigns for smart cities on the public's perceptions. Nonetheless, the important outcomes revealed the projection of promotional strategy in smart city potentially shift public attention from the socio-economic issues and unsatisfied policies. The similar pattern goes to the research from Vanjare et al. [12] which deploys a native bayes model to detect the problem raised by the user in twitter. The approach of this research confirmed the authorities to do a real-time examination of the issues arise in the society, enabling the quick actions and effective decision making through the response awareness. The suggestion also provides the method usage to be implemented for anticipating future understanding before the problem occurs. Furthermore, the application of Twitter in predicting the social conditions assures the benefit of crowdsourcing free data to be analyzed. Alhijawi and Awajan [13] collected a dataset of TweetAMovie from IMDb and Twitter to predict user satisfaction and popularity from the movie aired. This research has been successfully achieved the goals set with developed two types of prediction model as specifically rating prediction (RPM) and temporal popularity prediction models (T-PPM) in high accuracy. Moreover, with the technology improvement along the way, the data collecting method based on twitter is suggested to be integrated with other system such as information retrieval and recommender systems.
Smart This research aims to gather valuable information as a decision-making for the future application of smart city. The context of centralization and decentralization was designed to find the appropriate keywords from the objective users on social media. Our research implication procedure is distinguished into three main categories as the data collection, data processing, and classification, and summarized in the geo-location data visualization. Thus, the goal set in this research identifies the user-based analysis as the support decision on improving and predicting the future people satisfaction toward the decision-making obtained from Twitter as selected social networks.
Fig. 1. Research Process
In this stage, all of the user's data will be retrieved from Twitter throughout the application programming interface (API). This research will incorporate scrapping process using Python's library of BeautifulSoup that accessible and opensource provided by Jeddi [14]. In this process the research will scrape all of the information regarding the keywords assigned, given dates, list of works, and location. Our keywords set to any issues that raising on the society that includes either "Centralization" or "Decentralization." Therefore, the scraped information will be organized in the .CSV format for adjustable content with following the continuous steps designed. The tweets will also be managed in following the regional distribution yet the particular id of the city and country. Moreover, and advanced analysis for the sentimental purposes' interpretation will also be prepared.
Besides to handle the noisy data captured from the scraping process, this stage will also implement translation process of developing unstructured data into structured data [15, 16] as follows:
TFIDF will be utilized with aim of measuring the important words in the whole tweets produced in the previous stage. Hence this process will convert the textual data into the numerical value for the comprehensive classification on calculating the weight of word or feature [16]. The projecting illustration can be seen in Fig. 2. Which will also be based on the equation 1 below.
(1)
The result obtained from this weighting process calculated from the TF and IDF values which will serve a low result when the words frequently appear on each document and collection integrated.
![Fig. 2. Display of term-document matrix [17]](https://prod-files-secure.s3.us-west-2.amazonaws.com/87ab0b7d-a0ee-4393-a835-761fb82082d9/8ecbc1a5-0210-44fe-8bca-8a4135dd9f5b/Untitled.png)
Fig. 2. Display of term-document matrix [17]
The weighted-words resulted in the previous process will be further validated in the K-fold cross validation which distinguish the training dataset and testing dataset with K value iterations. The value will be visualized such as the following cross validation illustration that has been folded for 10 times.
![Fig. 3. 10-fold cross validation illustration [16]](https://prod-files-secure.s3.us-west-2.amazonaws.com/87ab0b7d-a0ee-4393-a835-761fb82082d9/f175afb4-4d6d-44b5-a867-16002747ecdd/Untitled.png)
Fig. 3. 10-fold cross validation illustration [16]
The This stage set the goal to classify the keyword of "Centralization" and "Decentralization" into the group with implementing Support Vector Machine (SVM), Multinomial Naive Bayes (MNB), and Linear Regression (LR) methods.
The collected tweet data obtained from Twitter, when it is already being processed and analyzed will be integrated with the context-based analysis of centralization and decentralization keywords assigned earlier. Moreover, along with the high frequency of the risen issues, this research will classify the words based on the location where the tweets intend to post and yet visualize in Kepler.gl. Kepler is an open source geospatial analysis tool to sketch the location-based data and deliver the data-driven decisions [18]. Besides this research also expect some broader understanding of the objective data distribution in the different city with the help of Twitter as a social network. This research proposes to obtain data from any cities in Taiwan as a preliminary study, thus wider analysis will be implemented simultaneously together with the prospective result achieved.
![Fig. 4. Geo-location-based data visualization [18]](https://prod-files-secure.s3.us-west-2.amazonaws.com/87ab0b7d-a0ee-4393-a835-761fb82082d9/b2200f4f-9b77-4437-b5fa-b0043c5c2fcd/Untitled.png)
Fig. 4. Geo-location-based data visualization [18]
This project mainly focuses on evaluating the advanced user's data with crowdsourcing mechanism from Twitter. The social network considered as an objective user viewpoint regarding the open-source platform and API provided. The research conclusion wrapped up with different measurement incorporated in TF-IDF as word weighting measurement, K-fold cross validation for validating the dataset, and SVM, MNB, LR as predictive methods classification. Nonetheless, the data visualization analysis is also incorporated to address the geo-location's source to strengthen the user's demands, issues' argument on the society. Furthermore, the future implementation of this study is looking for long term analysis and trends' prediction to improve the result better. Besides, another social network needs to be considered since the different users and cultures can be influence the popularity of the platform in many places.
[1] H. M. Alshuwaikhat, Y. A. Aina, and L. Binsaedan, "Analysis of the implementation of urban computing in smart cities: A framework for the transformation of Saudi cities," (in en), Heliyon, vol. 8, no. 10, p. e11138, Oct 2022, doi: 10.1016/j.heliyon.2022.e11138.
[2] D. Zeng, Y. Tim, J. Yu, and W. Liu, "Actualizing big data analytics for smart cities: A cascading affordance study," (in English), International Journal of Information Management, vol. 54, p. 102156, Oct 2020, doi: 10.1016/j.ijinfomgt.2020.102156.
[3] N. Steils, S. Hanine, H. Rochdane, and S. Hamdani, "Urban crowdsourcing: Stakeholder selection and dynamic knowledge flows in high and low complexity projects," (in English), Industrial Marketing Management, vol. 94, pp. 164-173, Apr 2021, doi: 10.1016/j.indmarman.2021.02.011.
[4] Y. Wang, S. D. Gao, N. Li, and S. Y. Yu, "Crowdsourcing the perceived urban built environment via social media: The case of underutilized land," (in English), Advanced Engineering Informatics, vol. 50, p. 101371, Oct 2021, doi: 10.1016/j.aei.2021.101371.
[5] Y. Zhang, M. Shirakawa, Y. Wang, Z. Li, and T. Hara, "Twitter-aided decision making: a review of recent developments," Appl Intell (Dordr), vol. 52, no. 12, pp. 13839-13854, 2022/09/01 2022, doi: 10.1007/s10489-022-03241-9.
[6] I. Shahrour and X. Y. Xie, "Role of Internet of Things (IoT) and Crowdsourcing in Smart City Projects," (in English), Smart Cities, vol. 4, no. 4, pp. 1276-1292, Dec 2021, doi: 10.3390/smartcities4040068.
[7] J. I. Castillo-Manzano, M. Castro-Nuño, L. Lopez-Valpuesta, and J. Boby, "Centralization versus decentralization of traffic law enforcement governance in Spain," Regional Studies, vol. 56, no. 11, pp. 1976-1988, 2022/11/02 2022, doi: 10.1080/00343404.2021.2020743.
[8] R. A. Mushkani and H. Ono, "Urban planning, political system, and public participation in a century of urbanization: Kabul, Afghanistan," Cogent Social Sciences, vol. 8, no. 1, p. 2045452, 2022/12/31 2022, doi: 10.1080/23311886.2022.2045452.
[9] S. P. Ramu et al., "Federated learning enabled digital twins for smart cities: Concepts, recent advances, and future directions," Sustainable Cities and Society, vol. 79, p. 103663, 2022/04/01/ 2022, doi: https://doi.org/10.1016/j.scs.2021.103663.
[10] A. J. Cheng, B. Tarroja, B. Shaffer, and S. Samuelsen, "Comparing the emissions benefits of centralized vs. decentralized electric vehicle smart charging approaches: A case study of the year 2030 California electric grid," Journal of Power Sources, vol. 401, pp. 175-185, 2018/10/15/ 2018, doi: https://doi.org/10.1016/j.jpowsour.2018.08.092.
[11] R. N. Arku, A. Buttazzoni, K. Agyapon-Ntra, and E. Bandauko, "Highlighting smart city mirages in public perceptions: A Twitter sentiment analysis of four African smart city projects," Cities, vol. 130, p. 103857, 2022/11/01/ 2022, doi: 10.1016/j.cities.2022.103857.
[12] N. Vanjare, N. Sarodi, R. Tantry, A. Koshe, and R. RB, "Real-Time Citizen Problem Detection From Twitter Data Using Naive Bayes Classifier," Available at SSRN 4097217, 2022.
[13] B. Alhijawi and A. Awajan, "Prediction of Movie Success Using Twitter Temporal Mining," in Proceedings of Sixth International Congress on Information and Communication Technology, Singapore, X.-S. Yang, S. Sherratt, N. Dey, and A. Joshi, Eds., 2022// 2022: Springer Singapore, pp. 105-116.
[14] A simple and unlimited Twitter scraper with python. (2023). Accessed: 2023-01-04 11:52:32. [Online]. Available: https://github.com/Altimis/Scweet.git
[15] K. Purwandari, J. W. C. Sigalingging, T. W. Cenggoro, and B. Pardamean, "Multi-class Weather Forecasting from Twitter Using Machine Learning Aprroaches," Procedia Computer Science, vol. 179, pp. 47-54, 2021/01/01/ 2021, doi: 10.1016/j.procs.2020.12.006.
[16] J. A. Septian, T. M. Fachrudin, and A. Nugroho, "Analisis Sentimen Pengguna Twitter Terhadap Polemik Persepakbolaan Indonesia Menggunakan Pembobotan TF-IDF dan K-Nearest Neighbor," Journal of Intelligent System and Computation, vol. 1, no. 1, pp. 43-49, 08/21 2019, doi: 10.52985/insyst.v1i1.36.
[17] N. Kaushik and M. K. Bhatia, "Twitter Sentiment Analysis Using K-means and Hierarchical Clustering on COVID Pandemic," in International Conference on Innovative Computing and Communications, Singapore, A. Khanna, D. Gupta, S. Bhattacharyya, A. E. Hassanien, S. Anand, and A. Jaiswal, Eds., 2022// 2022: Springer Singapore, pp. 757-769.
[18] Kepler.gl. "kepler.gl." https://kepler.gl (accessed 2023-01-04.
<aside> <img src="https://super-static-assets.s3.amazonaws.com/bb40d70c-1d06-4a05-8516-78a1cd5352cc/images/5ad7c2b4-49b3-4b55-bcf1-7be445ca0a1f.png" alt="https://super-static-assets.s3.amazonaws.com/bb40d70c-1d06-4a05-8516-78a1cd5352cc/images/5ad7c2b4-49b3-4b55-bcf1-7be445ca0a1f.png" width="40px" /> Categories
</aside>
<aside> <img src="https://super-static-assets.s3.amazonaws.com/bb40d70c-1d06-4a05-8516-78a1cd5352cc/images/5ad7c2b4-49b3-4b55-bcf1-7be445ca0a1f.png" alt="https://super-static-assets.s3.amazonaws.com/bb40d70c-1d06-4a05-8516-78a1cd5352cc/images/5ad7c2b4-49b3-4b55-bcf1-7be445ca0a1f.png" width="40px" /> Articles
</aside>
