By Dr. Hans Rocha IJzerman and Prakhar Srivastava

Some people feel deeply connected to their neighbourhood—invested, engaged, and truly at home—while others, even just doors apart, feel isolated or disconnected. Why does this happen?

At Annecy Behavioral Science Lab (ABSL), in collaboration with Neighbourly Lab , we explored this important question within London’s Borough of Camden. Our aim was clear: identify the factors shaping belonging, uncover who feels disconnected, and provide actionable insights to strengthen community bonds.

Belonging significantly impacts our mental health, civic involvement, and trust in local institutions. Yet public policy often overlooks it, considering it intangible or too complex to address. Our research seeks to make belonging tangible and actionable through advanced data analysis.

Our Approach: Harnessing Machine Learning

We analyzed data from the Survey of Londoners (2021/22) using a machine learning method called Conditional Random Forests (CRF). This approach enabled us to examine hundreds of potential predictors of belonging, from housing conditions and caregiving roles to neighbour contact and public space use. It also helped us understand how these factors interact, revealing patterns that more traditional analyses might miss.

Important note: While CRF highlights strong correlations, it does not establish causation. Findings offer valuable guidance but do not guarantee that modifying a factor will enhance belonging.

Four Essential Predictors of Belonging

Our analysis revealed four key predictors of belonging that were not only the strongest individual factors but also robust across different groups and consistent in their effects:

1. Working Status and Caregiving Roles

People who are employed full- or part-time, were more likely to feel a sense of local connection. These roles often anchor individuals in daily routines that facilitate community contact and provide structure. 

Working status and caregiving. This graph shows how employment status affects the sense of belonging in Camden. Working full-time has the strongest association with belonging, followed by working part-time and assumed working. Those not working due to long-term sickness or disability show the lowest belonging scores, with unemployed residents also reporting lower scores. Interestingly, people looking after households/children and retired individuals show moderate belonging levels. The data suggests that active participation in the workforce, particularly full-time employment, creates significant community connections.

2. Housing Tenure and Residential Stability

Stability played a vital role. Residents who had lived locally for longer, and those with more secure housing arrangements (like owning a home or holding a long lease), reported stronger feelings of belonging. 

Housing Tenure and Residential Stability Graphs. These graphs illustrate how housing stability influences belonging in Camden. The left graph shows belonging peaks after 5-10 years of residency but dips for both newcomers (less than 1 year) and lifelong residents. The right graph reveals that housing tenure relates to belonging, with residents from the Local Authority and homeowners reporting the highest scores, while private renters show the lowest. Together, these visualizations demonstrate how stable, secure housing arrangements foster stronger community connections.

3. Regular Neighbour Interaction

Even simple, consistent interactions—like a weekly greeting—were strongly associated with belonging. What mattered most wasn’t necessarily deep friendship, but the regularity of seeing familiar faces and exchanging small gestures of recognition, which builds a foundation of trust.

The following graph visualizes the relationship between frequency of neighbour contact and belonging scores.

Neighbor Interaction Graph. This graph reveals the complex relationship between frequency of neighbor contact and belonging. Belonging scores are highest with regular weekly contact (2-3 times a week), while monthly contact shows the lowest scores—even lower than no contact at all (we are unsure why this pattern emerges. The data suggests frequent, consistent interaction is key to community connection, but occasional or rare contact might actually feel more isolating than none.

4. Public Space Use

This emerged as a powerful, often underappreciated driver. Residents who spent time in local parks, markets, libraries, and other shared spaces felt more connected. These spaces provide not just amenities but opportunities for incidental interaction and shared presence that strengthen community bonds.

The following graph shows how frequent use of public parks is associated with stronger feelings of belonging.

Public Space Use Graph. This graph shows how park usage relates to belonging in Camden. Weekly park visits correspond with the highest belonging scores, significantly higher than for those who never visit parks. Again, for some reason visiting “around once a month” shows the lowest belonging score, while less frequent visits (“less than once a month”) show higher scores. The data suggests regular engagement with public spaces strengthens community connection, though the relationship isn’t strictly linear.

Who Feels Left Out? Identifying Camden’s Disconnection Hotspots

Our data doesn’t just tell us who feels connected—it clearly reveals who doesn’t. By examining the inverse patterns of our key predictors, we’ve identified specific groups experiencing significant belonging deficits in Camden:

• Private renters emerge as particularly vulnerable, especially those caught in cycles of frequent moves due to short-term leases or rising costs. Unlike homeowners or social housing tenants who can put down roots, these residents live with perpetual uncertainty that undermines community investment.
• Newer residents face substantial belonging challenges during their first three years in Camden. Without established routines or local connections, they often remain peripheral to community life, especially if they lack institutional anchors through work, school, or family.
• Residents with minimal neighbor contact report some of the lowest belonging scores in our dataset. This includes people in buildings with limited common areas, streets with few natural gathering points, or housing layouts that discourage interaction. Their physical proximity to others doesn’t translate into meaningful connection.
• Public space non-users miss crucial opportunities for community integration. Whether due to mobility challenges, safety concerns, time constraints, or simple lack of awareness, residents who rarely visit parks, libraries, or community centers show markedly lower belonging.
• Those in precarious housing situations beyond private renting also struggle significantly. This includes residents in temporary accommodation, those facing potential displacement through regeneration, and people in overcrowded housing seeking transfers. The psychological impact of housing insecurity appears to extend into community disconnection.
• Working caregivers facing severe time pressure present a paradoxical case. While caregiving roles generally correlate with higher belonging, those juggling excessive responsibilities without adequate support systems find their community engagement severely restricted.

These patterns of disconnection reveal an important truth: belonging deficits don’t exist in isolation. They often cluster and compound each other, creating a cycle that’s difficult to break without intervention. But this interconnected nature of disconnection also reveals something encouraging about belonging itself—it emerges from multiple pathways, offering several potential entry points for fostering community connection.

Belonging Emerges from Multiple Factors, Not Just One

Belonging does not hinge on a single factor. Our research clearly demonstrates that it arises from the interplay between various elements of community life.

For example, residents who combine regular neighbour interactions with frequent use of public spaces report notably stronger connections than those experiencing just one or neither. The CRF heatmaps showed exactly this kind of synergistic effect – the highest belonging scores occurred in cases where community behaviors coincided (e.g., high public space use plus high neighbor contact). Understanding these combined effects is crucial for designing effective community-building interventions.

Heatmap for Public Spaces and Contact with Neighbors. This heatmap visualizes how neighbor interaction and park usage work together to influence belonging. The brightest yellow areas (highest belonging scores) appear when frequent neighbor contact (daily/weekly) combines with regular park visits (3+ times weekly). The darkest blue areas (lowest scores) occur with monthly neighbor contact regardless of park usage. The visualization reveals these factors amplify each other’s effects—frequent engagement in both creates significantly stronger belonging than either alone, highlighting the importance of multiple connection pathways in community building.

Other interaction insights suggest different life circumstances intersect. Working status can moderate the effect of neighbor interactions, for instance. A newcomer to Camden who is also an active caregiver might integrate faster and feel more belonging than a newcomer without those social ties. Belonging, in this sense, is more than a product of identity or background. It grows out of the intersections of social, personal, and place-based experiences.

Real-Life Inspiration: The Hypervoisins Movement in Paris

The data gave us clarity, but stories help us imagine what belonging looks like in practice. One model we kept returning to is Hypervoisins—a grassroots movement in Paris that has reimagined how neighbours connect. Launched in 2017, this initiative transformed about 50 ordinary city streets (encompassing 15,000 residents) into a vibrant village-like community through simple acts of daily conviviality.

It began with a 215-metre-long street banquet and a simple idea: say bonjour to the people around you. That one event grew into something larger—a rhythm of weekly neighbour-led meetups, outdoor movie nights, garden projects, WhatsApp groups, and shared moments in public space. Today, more than 2,000 people participate every week. As founder Patrick Bernard describes it, they aimed to “create the atmosphere of a village in an urban space,” and it’s working.

The impact has been profound: stronger social bonds, quicker mobilisation during emergencies like COVID, and a neighbourhood that feels more like a village—despite being in the heart of a busy city. The founders didn’t rely on expensive technology or infrastructure. They simply created regular opportunities for people to show up, contribute, and connect.

What makes Hypervoisins so inspiring is that it activates the exact drivers we identified in Camden’s data—caregiving routines, neighbour interaction, public space use—and ties them together in a low-cost, high-trust model that any community can adapt.

Camden’s Opportunity: Bringing Insights to Life

Camden, with its diversity, energy, and strong tradition of civic engagement, is already pioneering community-building approaches that align remarkably well with our research findings. As part of their Neighbourhoods programme, they have worked with community leaders to establish Kentish Town Connects, which aims to collectively prioritise, support and deliver community-led projects for the benefit of residents, organisations and everyone within Kentish Town. This builds on the success of the One Kilburn project, which is a network of Kilburn community organisations, residents, businesses, local partners and stakeholders who have come together with a focussed vision to support the whole of Kilburn. 

Both of these projects represent a different mode of work for a locality. Rather than engaging with residents based on a pre-set agenda, they focus on long term relationship building and open inquiry. The mantra of ‘spending time with people, in place’ has shown dedication to the place and emphasised that when working with communities in this way you can only move at the speed of trust. This approach—letting connection develop organically rather than forcing it through predetermined programs—mirrors the spontaneous community-building seen in the Hypervoisins movement, while adding the institutional support and inclusivity focus that Camden excels at providing.

From Parisian Streets to Camden Communities: A Path Forward

The Hypervoisins movement in Paris offers a complementary blueprint that aligns perfectly with both our data findings and Camden’s existing initiatives. By activating the exact drivers we identified—neighbor interaction, public space use, community routines—the Parisian model transformed disconnected streets into connected villages through simple, low-cost interventions. 

What makes this approach particularly promising for Camden is how naturally it could enhance and expand the borough’s existing work. The regular neighbor-led meetups, outdoor community events, and WhatsApp groups of Hypervoisins could complement Camden’s institutional support structure, creating a hybrid model that leverages both grassroots energy and council resources.

Practical Steps: A Camden Model for Belonging 

Building on Camden’s impressive foundation, we envision a uniquely Camden approach to community connection that integrates insights from our research with lessons from both Kentish Town Connects and Hypervoisins: 

1. Scale the “Spending Time with People, in Place” Model: Expand the successful approaches from Kentish Town Connects and One Kilburn to additional neighborhoods, maintaining their emphasis on relationship-building and open inquiry while incorporating the regularity of interaction seen in Hypervoisins. By honoring the principle that communities can “only move at the speed of trust,” this approach ensures sustainable connection rather than superficial engagement. The listening campaigns, stakeholder interviews, and community workshops already happening in these initiatives are precisely the qualitative approaches needed to understand belonging at a deeper level.
2. Hyper-Local Focus on Disconnection Hotspots: Target interventions toward the specific groups our research identified as experiencing belonging deficits—private renters, newcomers, and time-pressed working caregivers. For example, create “Welcome Teams” for streets with high turnover, partner with estate agents and landlords on newcomer packets, and develop flexible engagement opportunities for residents with limited time. 
3. Activate Public Spaces as Connection Hubs: Draw inspiration from Hypervoisins’ street banquets and outdoor gatherings while leveraging Camden’s existing public realm strategy. Support resident-led initiatives to transform parks, streets, and common areas into regular gathering spaces, focusing particularly on areas where neighbor contact is minimal. 
4. Foster Connection Across Housing Tenure: Develop targeted approaches for private renters that acknowledge their mobility while still creating opportunities for connection. “Renters’ Collective” events could combine practical support with community building, while mixed-tenure street events could help integrate residents across housing types. 
5. Create Belonging Rituals and Routines: Establish weekly, predictable opportunities for connection in each neighborhood—whether Sunday picnics, Thursday gatherings, or Tuesday community dinners. The regularity creates touchpoints that residents can build into their routines, addressing the belonging challenges faced by time-pressed residents. 

Measuring Impact While Building Trust 

Camden’s commitment to addressing inequalities already incorporates sophisticated data practices. Building on this foundation, a phased implementation approach offers a way to measure impact while ensuring all communities eventually receive support. 

Programs introduced to different streets or neighborhoods in staggered phases create natural comparison groups at each stage, generating evidence about what works while honoring the principle that all communities deserve support. 

By developing indices that capture the multidimensional nature of community connection—drawing on the four key predictors identified in our research—Camden could track belonging as a metric for community wellbeing. This approach would connect Camden’s impressive work to the growing global conversation around social connection in public policy. 

Co-Creation as the Camden Way 

The We Make Camden Kit exemplifies the co-creation philosophy that must remain central to successful community building. This approach creates platforms where residents lead while local authorities provide strategic support, resources, and removal of barriers. By combining the institutional support of Camden Council with the neighbor-to-neighbor connection model of Hypervoisins, Camden could pioneer a hybrid approach to belonging that draws on the best of both worlds. 

Looking ahead, Camden is well-positioned to become a borough where newcomers quickly find connection, where streets serve as social spaces, where neighbors of different backgrounds find common ground, and where belonging flourishes regardless of housing tenure or family structure. The data points to this vision’s potential, Camden’s own initiatives demonstrate what it already looks like in practice, and the Hypervoisins blueprint offers a complementary path forward. Together, these elements create an opportunity for Camden to continue writing its own success story of connection—one street, one park, one neighborhood at a time.