An Innovative Application of Social Media Analytics for Smart City Planners

Shah J. Miah

Newcastle Business School, University of Newcastle, Australia.

This is a translation of an article which first appeared in JASIST.

Social media data hold various types of people’s activity and engagement details online. The details of people’s daily activities in a particular region can be used to reproduce priceless knowledge for smart city planners. The new knowledge may subsequently be utilized for improving economic activities and conducting sustainable environmental impacts, responding to improve the quality of human life. While data-production in social media provides opportunities for developing evidence-based insights into human daily activities, it is important to develop a new analytical method to address such needs.

The motivation of the study

In city living, quality of life, in a way, refers to smartly meeting human preferences and habitual activities that result in affective well-being. This could be assisted through enabling options for increasing support and socialization for people in their living space (De Guimaraes et al. 2020). Population growth in urban areas has created many issues that affect various strategic decisions of government authorities. As a part of smart city governing, it is important to reveal human activities and their preferences for planning, specifically for better planning to meet upcoming challenges (De Guimaraes et al. 2020). For instance, planners of smart cities are encouraged to explore options of healthy economic activities which reduce the environmental burden while they work for improving quality of life. Therefore, smart city governance makes use of new technologies to develop innovative practices (Meijer et al., 2016). The innovation may be putting enormous volumes of data at the core of smart city management that can potentially improve city services, such as assistance to building shopping complexes, public transportation planning, and providing public information. In this context, data analytics techniques provide new avenues to address the demands of capturing and processing different forms of human-generated data online.

Social media analytics

Venue Referenced data that are generated in social media can be viewed as a new form of location-referenced data, commonly used for location analysis (Lee et al. 2019). The advantage of this data is its potential for revealing insights into human activity. This is a type of big data that preserves elements about a citizen who visited venues such as restaurants, shopping centers, and railway hubs in a city destination. These destinations can be used to infer their various daily activities (e.g., their dining, shopping, and traveling). Existing approaches have been used for processing and analyzing location data (Huang et al., 2019) for giving us insights on the movement of citizen. But, this type of approach does not offer insights into the activity context of venues (e.g., what people are doing there). Thus, it is imperative to develop a new method for generating new insights as a secondary source of evidence, which is convenient for decision-makers’ records.

As a social-media analytics solution, a new approach (Miah, Vu, & Damminda, 2021) was introduced to utilize venue-referenced big datasets. The approach adopted topic-modeling techniques, particularly Latent Dirichlet Allocation (Blei et al, 2003), for producing evidence-based insights into human daily activities. In this context, the proposed approach goes beyond the previous approaches for tracking human activities using social-media data for investigating diversified perspectives. For instance, various studies have studied urban issues of developing predictors to understand human mobility and activity patterns for determining traffic congestion. The proposed approach (see Figure 1) is represented as a conceptual framework, which comprises three major components:

1. Venue Referenced Data Collection: this component captures activity data of users on venue-referenced social-media platforms.
2. Daily Activity Modeling: this component works with metadata of the collected data preprocessing, where venue history records are converted into a bag of words representation for input into a topic modeling for modeling of daily activity.
3. Activity Analysis:  this component is responsible for analyzing records to understand the major themes of the daily activities. Further details of each component are described in Miah et al. (2021). Potential example outcome of the proposed approach is shown in Figure 2.

Figure 1: Proposed Analytic Approach (Adopted from Miah et al. 2021)

Figure 2: Example Venue Check-Locations in a country

Ending note

A new social-media analytics approach is useful for revealing evidence-based insights into human daily activities. The results showcase how the proposed method could address a potential issue of decision support (e.g., for smart city planners or professional transport planners for sustainable city growth). The developed method adopted a number of data analytics and topic modeling for revealing insights. This method is unique that goes beyond the existing methods used for analyzing textual data. Within the context of smart city management, our focus is centered primarily at the individual level phenomenon that for the first time examines the interplay of information technology and systems and individuals’ daily insights in the current digitally enabled society. 

References

Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet allocation. Journal of Machine Learning Research, 3, 993–1022.

De Guimaraes, J.C.F., Severo, E.A., Junior, L.A.F. Da Costa, W.P.L.B. & Salmoria (2020). Governance and quality of life in smart cities: Towards sustainable development goals, Journal of Cleaner Production, 253, 119926.

Miah, S.J., Vu, HQ., & Alahakoon, D. (2021). A social media analytics perspective for human-oriented smart city planning and management, Journal of the Association for Information Science and Technology, URL: https://asistdl.onlinelibrary.wiley.com/doi/abs/10.1002/asi.24550

Huang, W., Xu, S., Yan, Y., Zipf, A. (2019). An exploration of the interaction between urban human activities and daily traffic conditions: A case study of Toronto, Canada, Cities, 84, 8-12.

Lee, K.S., Yu, S.Y., Eom, J.K., Song, J., & Min, J.H. (2019) Urban spatiotemporal analysis using mobile phone data: Case study of medium- and large-sized Korean cities, Habitat International, 73, 6–15.

Meijer, A., Gil-Garcia, J. R., & Bolívar, M. P. R. (2016). Smart city research: Contextual conditions, governance models, and public value assessment, Social Science Computer Review, 34(6), 647–656.