"For You" City
Written by: Athens (25-I1)
Designed by: Jia Jia (26-A3)
"For You" City (FYC)
The 2011 General Election (GE2011), is often referred to as the ‘watershed election’. It was the first time since independence that the ruling People’s Action Party (PAP) faced a serious threat to its dominance. The results for GE2011 signalled a shortfall in support among younger voters.
The looming reality of a widening disconnect between youths and policymakers threaten to derail the furture of our built environment. The urban policies created for the city may fail if they do not meet the actual needs of the people using them, as inclusion is a critical prerequisite for a successful implementation of urban solutions. Without an accurate understanding and representation of the youth voice, the facilities that the government builds might risk becoming obsolete if the youth do not use them. It could also significantly affect the way people relate to their community, as there is a lack of stewardship in the public spaces that the youths had no part in throughout the implementation process. Events like ‘Our Singapore Conversation’ and ‘Forward Singapore’, are created to allow youths to voice out all of their concerns and share some of their aspirations and their goals for themselves and the country. The introduction of ‘Youth Panels’ provided youths a direct seat at the policy-making table. According to Sherry Arnstein’s Ladder of Citizen Participation, most youth engagement organisations are merely tokenistic, though to differing degrees. Hence, the bond of trust between the authorities and the younger generations will inevitably be fractured if the government persists in using tokenistic ways to engage the youth.
This article is mainly focused on the premise that youths’ relationship with the urban spaces is increasingly reconfigured by social media. Hence, the relationship, specifically emotionally, that youths develop with urban environments can be understood through the integration of both spatial experience and algorithmic mediation.
Video recommender systems, particularly short-form videos on TikTok and Instagram Reels, actively shape perceptions towards different locations through the amplification of videos via repeated exposure, visual aesthetics and engagement metrics.
Hence, algorithmic recommendation systems are positioned as the active agents responsible for shaping youths’ emotional landscapes and spatial behaviours and this article seeks to find out how digital media influences the lived experience of urban space.
Urban Theory
Theory of Place Attachment
Place attachment explains how individuals develop emotional bonds with places. Place attachment often operates through broader affective and functional connections to the physical environment itself. Place attachment mediates between the neighbourhood environment and individuals’ emotional experiences of place.
Based on foundational work in environmental psychology, geography and sociology, place attachment is commonly defined as “… a set of feelings about a geographic location that emotionally binds a person to that place”. There is a set of frameworks that distinguishes place attachment from overlapping related notions such as sense of place or place belonging. Firstly, place identity refers to the symbolic and emotional connections through which places become integral to an individual’s sense of self. It often reflects belonging and personal history. On the other hand, place dependence captures the functional and goal-oriented value of a place, which evaluates the extent to which the place supports individuals’ daily needs and activities relative to alternate locations.
Prior research suggests that environmental and residential characteristics influence place attachment. The characteristics of an individual’s living environment shape the degree to which individuals feel emotionally and functionally connected to their surroundings.
This literature suggests that place attachment offers added explanatory value in mediating the relationship between environmental conditions and subjective well-being. Place attachment complements social integration perspectives and highlights the importance of both physical and emotional connections to a place in shaping experiences.
b. Urban Features and Impact on Emotional Response
Existing research shows how specific sensory compositions of nature can influence the liveliness of human activities and psychological well-being in the urban context. To achieve this, the researchers conducted a case study at the Jiangjia Art Garden in Chengdu, China.
The results that were obtained indicated a significant positive correlation between nature-dominated environments and both human activity liveliness and emotional well-being. It concluded that emotional responses are deeply correlated with the behavioural patterns observed in naturalised settings. Hence, the design of high-quality urban environments must leverage both sight and sound to promote human physical and mental health.
Youth Studies
Young People’s Participation in Public Space
Youth are becoming more significant as a demographic within rapidly urbanising contexts, particularly in the Global South.
Existing research emphasises that participation in public space is central to young people’s social, emotional and civic development. By interacting with urban environments, youth develop social relationships and have a sense of responsibility, which is often mediated through place attachment. Place attachment is particularly important during youth due to rapid psychosocial development. Participatory engagement in the planning and decision-making process has been shown to enhance young people’s sense of belonging, empowerment and civic identity, while also contributing positively to community outcomes. The impacts of youth participation in planning are often categorised and regarded as social and physical, benefiting both young participants and the wider community in reciprocal ways.
However, young people remain systematically marginalised within formal planning processes. Assumptions about youth dependency on adult supervision and the limited institutional mechanisms for meaningful engagement constrain their participation. Effective youth participation is further hindered by practical and ethical challenges, which include difficulties in sustaining engagement. Common participatory formats such as youth councils often involve a small number and feature an underrepresented subset of young people and they rarely grant substantive influence over decision-making. However, when youth are handed with the opportunity to influence, the results are transformative. In October 2023, PlayPan took over the Peace Centre in Singapore. Before the slated demolition of the Peace Centre, the mall was handed over to youths who transformed it into a space for non-profit and social engagement activities. This proved that once youths are treated as co-creators, the landscape of Singapore can transform into a thriving community.
b. Placemaking
In response to these limitations, placemaking has emerged as an alternative approach to youth participation. Placemaking focuses on informal, collaborative and place-based processes that reduce bureaucratic barriers and prioritise experiential hands-on engagement. By positioning young people as experts in their own environments, placemaking gives youth a chance to shift from passive users of the space to contributors. This approach allows the transformation of public spaces into meaningful places through attention to different aspects of dimensions, including but not limited to the social, cultural, economic and political, while also fostering identity expression and social connection.
Algorithms and Mediation of Experience
Collaborative Filtering
Collaborative Filtering (CF) is a technique used in recommender systems, intending to find and provide specific and personalised recommendations through the leverage of users’ tastes and preferences and grouping users with the same tastes and behaviours.
CF operates under the principle that users with similar interests are potentially rating items similarly.
b. Representation Learning
Representation learning generates personalised representations for each user and item, ultimately playing a salient role in collaborative filtering. These personalised representations accurately reflect the interests of each user and characteristics of each item, which leads to a heightened accuracy and effectiveness in CF-based models. Representation learning is the key asset that helps models capture the underlying relationship between users and items accurately, which results in more accurate recommendations.
c. Current State of Video Recommender System (VRS)
Video Recommender Systems (VRS) have become fundamental components of modern digital infrastructure by facilitating personalised content consumption across different fields. As digital video catalogues continue to grow in volume and complexity, the necessity for a sophisticated filtering mechanism continues to grow as well. This is to ensure that users can efficiently navigate the vast amounts of information available.
Today, platforms like Netflix and YouTube rely heavily on video recommender systems (VRS), with approximately two-thirds of Netflix’s streamed content originating from personalised suggestions.
In Figure 4, this approach recommends new items based on the features of the content of the video itself, rather than how other users behave. For instance, if a user likes video (V1), the system analyses its metadata to find and recommend video (V2), which share similar features.
Panel B outlines collaborative filtering, which shifts the focus toward social similarity. This mechanism identifies ‘neighbours’ or users who demonstrate shared preferences. When user A and user B both exhibit a preference for (V1), they are categorised together for having similar tastes. Consequently, the system recommends (V2) to user A based solely on the fact that user B liked it, operating on the premise that individuals with shared past interests will likely agree on their liking for future content.
Panel C showcases a hybrid recommendation, a more scalable variation of the collaborative approach. Rather than comparing users directly, this method calculates the similarity between items based on the patterns of interaction from the entire user base. When (V3) is recommended to user A because the collective data suggests that (V2) and (V3) are frequently enjoyed by the same cohort of viewers. Hence, this creates a behavioural map of item relationships that transcends simple metadata.
Lastly, panel D illustrates group recommendation. This sophisticated approach aggregates the multiple users into a collective profile to refine the recommendation engine’s accuracy. By analysing the intersection of group preferences and the multifaceted similarities between a network of item (V1), (V2) and (V3) the system is then able to provide a highly tailored suggestion. This model effectively mitigates individual data noise by leveraging the broader trends across a demographic or behavioural cluster.
d. Influence of Artificial Intelligence Algorithms on Worldview and Behaviour
Artificial Intelligence (AI) algorithms exert significant influence over individuals by shaping their beliefs and worldviews, which will influence their perceptions of social structures and political systems. ⁸ The algorithm affects an individual’s ability to make decisions, affects an individual’s beliefs and their relationships. This influence can be interpreted as a form of control operating through a cycle of restriction. Algorithms will deliver content based on user preferences, which will inadvertently confine and shape the user’s worldview.
An algorithm is a set of clear step-by-step instructions designed to perform a task by taking in data, processing and producing an output. AI uses the algorithm to increase efficiency and promote beneficence.
The primary mechanism of algorithmic control is the creation of an echo chamber, which is the process in which algorithms continuously supply information that reinforces an individual’s pre-existing beliefs. This endless stream of confirmation nudges and shifts beliefs towards a more rigid set of beliefs, exacerbating confirmation bias.
Experimental Investigation
70% of participants strongly agreed that they felt emotionally attached to New Bahru after exposure to the videos.
70% of participants visited the location after the experiment, which suggests that the experiment is a success in proving that recommendation videos play an important role in shaping the emotional landscape of a space and in shaping participants’ decision-making.
There is a strong positive linear correlation between the frequency of recommended video exposure and the percentage of participants who visited the recommended location. The visitation rate increases by a constant factor of 20% points for every 5-unit increase in video exposure.
Online Survey
TikTok is the most popular source of awareness of the participants’ chosen locations, with 66.7%.
The number of likes/views, visual aesthetic and repeated appearance of a recommended video largely influence participants’ interest in the location, with 68.4%.
66.7% of participants strongly agree that as they view more recommendation videos, they are more likely to visit the location.
Observational Comparison Case Study
Place | |||
Marina Bay Sands | Jewel Changi Airport | Our Tampines Hub | |
Footfall | 38,000,000 | 80,000,000 | 18,000,000 |
TikTok Location Posts | 439,000 | 120,000 | 9,886 |
TikTok Location Reviews | 27,900 | 19,600 | NA |
TikTok Location Rating | 4.9/5.0 | 4.9/5.0 | NA |
Google Maps Location Reviews | 61,213 | 91,419 | 13,174 |
Google Map Ratings | 4.7/5.0 | 4.8/5.0 | 4.5/5.0 |
Comparison Table between Marina Bay Sands, Jewel Changi Airport and Our Tampines Hub in Singapore on Footfall and Online Ratings
This table suggests a clear divergence between physical usage of urban spaces and their digital representation across platforms, particularly when comparing TikTok and Google Maps. Jewel Changi Airport records the highest annual footfall at approximately 80 million visitors, followed by Marina Bay Sands with 38 million and Our Tampines Hub with 18 million. These figures reflect the differing functional roles of the spaces, with Jewel serving as a major transport and leisure mode, Marina Bay Sands as a tourism and entertainment landmark and Our Tampines Hub operating primarily as a community-oriented facility.
Despite having a lower footfall than Jewel Changi Airport, Marina Bay Sands demonstrates a disproportionately strong digital presence on TikTok, accumulating approximately 439,000 location-tagged posts, far exceeding Jewel’s 120,200 posts and Our Tampnes Hub’s 9,886 posts. This underscores that TikTok visibility is not directly tied to the physical usage, but rather to the aesthetic and symbolic qualities of a space that encourage visual documentation and online sharing. Marina Bay Sands and Jewel Changi Airport both record a high TikTok location rating of 4.9 out of 5, alongside substantial numbers of TikTok reviews, whereas Our Tampines Hub lacks sufficient reviews and rating data, suggesting limited engagement on the platform.
On the other hand, Google Maps data presents a more utility-oriented perspective. Jewel Changi Airport receives the highest number of Google Maps reviews at over 91,000 and maintains a strong average rating of 4.8, reflecting consistent user satisfaction with its services and facilities. Marina Bay Sands then follows with over 61,000 reviews and a rating of 4.7, while Our Tampines Hub, despite its comparatively modest digital footprint, records over 13,000 reviews and a decent rating of 4.5. This is indicative of the community spaces being valued and frequently used by locals, even if they are less visible on visually driven social media platforms.
I have found out that there are various factors in which a recommendation video influences someone’s decision to visit the location recommended by the video. Such factors include the number of likes, the repeated appearance of the video on their feed, the visual aesthetic, the reviews or captions.
Engagement metrics such as the number of likes and views on a video act as social proof of the desirability of the location. It influences an individual’s decision to visit a location due to the powerful social and psychological cue that shapes perception and expectation of the value of the location. Individuals are more likely to trust highly liked content as popularity is often subconsciously equated with quality and authenticity. A video with high engagement is more likely to be amplified by the algorithm, increasing its visibility and frequency of exposure. Repeated exposure fosters familiarity, it becomes normalised that the location is a “must-visit”.
The visual aesthetic of a video also plays a role in influencing an individual’s decision to visit a location, as it shapes emotional responses and expectations of the experience that they will engage in. A visually appealing video constructs an idealised representation of a place, allowing viewers to anticipate how it might feel to be there. Through elements such as the lighting, adjustments of the colour, the composition of the video and camera movement, the video can make a location appear vibrant and enhance its perceived attractiveness. Aesthetic visuals play a role in heightening emotional engagement by eliciting different emotions. Viewers may come to associate the place with the positive internal states while watching due to the location being depicted in the videos.
Recommendation videos act as a form of interoceptive priming, allowing individuals to simulate the embodied experience of being in the location. This imagined embodiment reduces uncertainty and strengthens emotional attachment as the place is more familiar and accessible. Thus, the desire to visit stems from the motivation to re-experience or intensify the positive body sensations that were previously evoked through the viewing of the video.
Next, the reviews or captions frame how the visual content is interpreted. They play a role in shaping the user’s expectations and emotions. Captions and reviews act as interpretive cues that anchor the viewer’s experience. Descriptive language and anecdotal experiences not only provide supplementary information but also guide the viewer’s imagination, influencing how the location is internally visualised and anticipated.
However, there are several shortcomings that should be acknowledged when interpreting the findings.
The experimental component, which was designed to isolate the effects of algorithmic exposure, might simplify real-world social media use. Participants viewed curated feeds in a controlled and time-limited setting, which differs from the habitual, fragmented and prolonged nature of everyday scrolling. As a result, the emotional response and behavioural intentions captured may reflect short-term reactions rather than the cumulative effects of repeated algorithmic exposure over time, which will have a greater impact on the person.
Additionally, the results rely significantly on an individual’s capabilities to assess their own emotional perception, attachment and likelihood of visitation. Although a Likert-scale instrument allows for systematic comparison across participants, it may not fully capture the complexity of embodied experiences. The responses may be influenced by social desirability bias or participants’ limited ability to accurately articulate internal emotional and bodily states, specifically related to interoception.
CONCLUSION
This shows how algorithmic recommendation systems shape youths’ emotional responses to and their behavioural perception of the varying urban spaces in Singapore. By integrating survey data, the observational comparisons and the experimental investigation of algorithmic exposure, the research demonstrates that digital recommendation environments actively participate in producing spatial preferences.
Recommendation systems play a huge and critical role in shaping the emotional landscape of the space by making people feel more or less connected with the space through the promotion of activities to be done within the space, which would lead to an increase in visitation of the space. A prolonged engagement with the space makes it feel more vibrant and emotionally pleasant to community members, strengthening the ties and relationships within the community.
Ultimately, this issue directly impacts how we, as young people, experience our own city. Since youths are increasingly forming emotional attachments to spaces through algorithmic feeds, we must dig deep and reflect: are we visiting places curated for us by an app or are we really connecting with our community? Urban policy that relies solely on the traditional forms of consultation with youth groups risks overlooking how young people actually encounter, interpret and value urban spaces in their everyday lives. However, by understanding how digital spaces are a reflection of our real-world desires, we can change the narrative. By actively championing our authentic desires online and demanding to be heard can we then bridge the chasm between the state and the street – by ensuring that the future of Singapore is actively co-created by the generation that inherits it.
B, Q., DO, McKone, K. M. P., Hofman, S. J., Morrow, K. E., Brehm, M. V., Ladouceur, C. D., & Silk, J. S. (2025). Feeling socially connected to others is linked to better emotion regulation in adolescent girls’ daily lives: An ecological momentary assessment study. Journal of Research on Adolescence, 35(1), e70009. https://doi.org/10.1111/jora.70009
Blackmore, S. (2024, July 1). Why do we get a mental picture when reading? BBC Science Focus Magazine. https://www.sciencefocus.com/the-human-body/why-do-we-get-a-mental-picture-when-reading
Büssing, A. G., Wülfing, M., & Bruckermann, T. (2025). Emotions, connectedness to nature and sustainable behaviour: investigating the effects of students’ engagement in extracurricular activities with living animals. Journal of Biological Education, 1–18. https://doi.org/10.1080/00219266.2025.2452191
Chapman, M. (2025, October 14). Interoception: Scientists to create first atlas of Humans’ “Hidden Sixth Sense.” IFLScience. https://www.iflscience.com/humans-hidden-sixth-sense-to-be-mapped-following-142-million-prize-what-is-interoception-81155
Cordero-Vinueza, V. A., & Viese, V. (2025). Youth in Placemaking: Barriers and benefits. The Journal of Public Space, 10(1), 83–106. https://doi.org/10.32891/jps.v10i1.1808
Ellenbogen, N., & Trivic, Z. (2024). Dynamic place attachment in the context of displacement processes: The socio-ecological model. https://www.sciencedirect.com/science/article/pii/S0264275124000763
Fu, B., Wang, J., Liu, X., & Liu, Y. (2025). Can We Predict the Next Question? A Collaborative Filtering Approach to Modeling User Behavior. https://arxiv.org/abs/2511.12949
Fulmer, R. (2024). A Primer on How Artificial Intelligence Algorithms Control You. University of San Diego. https://digital.sandiego.edu/cgi/viewcontent.cgi?article=1098&context=tces
Gao, M., & Fang, C. (n.d.). Do urban park spatial features influence public emotional responses during jogging? Evidence from social media data. https://www.sciencedirect.com/science/article/abs/pii/S2213078025000106
Gutzeit, J., Dorsch, I., & Stock, W. (2021). Information Behavior on Video on Demand Services: User Motives and Their Selection Criteria for Content. https://www.mdpi.com/2078-2489/12/4/173
Jain, B., Pitsko, R., Drishti, A., & Farooque, M. (2025). Causally-Informed Reinforcement Learning for Adaptive Emotion-Aware Social Media Recommendation. In https://arxiv.org/pdf/2511.14768. https://arxiv.org/abs/2511.14768
Jewel Changi Airport reports record high footfall and sales amid steady growth in air travel. (2025, January 28). https://www.changiairport.com/en/corporate/our-media-hub/newsroom/press-releases.jewel-reports-record-high-footfall-and-sales.2025.all.html. https://www.changiairport.com/en/corporate/our-media-hub/newsroom/press-releases.jewel-reports-record-high-footfall-and-sales.2025.all.html
Jiang, J., Hong, Y., Li, W., & Li, D. (2022). A study on the impact of official promotion short videos on tourists’ destination decision-making in the post-epidemic era. Frontiers in Psychology, 13, 1015869. https://doi.org/10.3389/fpsyg.2022.1015869
Johanna Reads. (2022, June 23). Do you visualize what you read? 👀 [Video]. YouTube. https://www.youtube.com/watch?v=IM-JTJUQPU8
Kumar, P. (2025, October 16). Quality of Life's Influence on Place Attachment • Psychology Town. Psychology Town. https://psychology.town/environmental/quality-life-influence-place-attachment/#:~:text=Understanding%20place%20attachment%20%F0%9F%94%97,of%20belonging%20in%20the%20world.
Kumar, P. (2025, October 16). Understanding the Concept of Place in Environmental Psychology • Psychology Town. Psychology Town. https://psychology.town/environmental/environmental-psychology-concept-of-place/#:~:text=Place%20Attachment%20%F0%9F%94%97,even%20at%20a%20psychological%20level.
Lubos, S., Felfernig, A., & Tautschnig, M. (2023). An overview of video recommender systems: state-of-the-art and research issues. Frontiers in Big Data, 6, 1281614. https://doi.org/10.3389/fdata.2023.1281614
Occupi. (n.d.). Our Tampines Hub - Singapore | Premium Retail space | OCcUPI - Global Retail Property Marketplace. Occupi. https://getoccupi.com/malls/our-tampines-hub
Place attachment. (n.d.). Google Books. https://books.google.com.sg/books?hl=en&lr=&id=_-nvBwAAQBAJ&oi=fnd&pg=PA1&dq=research+about+place+attachment&ots=oUt8B3a_im&sig=4MRCAlJ1B3dUnhnjtBJg7m6iEgg#v=onepage&q=research%20about%20place%20attachment&f=false
Shen, Z., & Wu, D. (2025). Local Collaborative Filtering: A Collaborative Filtering Method that Utilizes Local Similarities among Users. In https://arxiv.org/pdf/2511.13166. https://arxiv.org/abs/2511.13166
Town, P. (2024). MARINA BAY SANDS ANNUAL REVIEW 2024. https://www.marinabaysands.com/content/dam/marinabaysands/company-information/MBS-Annual-Report-2024.pdf
Comments