Loneliness, inevitably associated with social interactions, has been shown to affect our mental and physical wellbeing (Henriksen et al., 2019). Its impact has increased in recent years due to the COVID-19 pandemic and the associated social changes (Kovacs et al., 2021). In the UK alone, nearly 20% of the population reports feeling lonely, which affects their sense of hopelessness and reinforces negative self-perception (Ypsilanti & Lazuras, 2022). To provide loneliness relief, wellbeing interventions such as mental health counselling, digital and offline peer support groups, and befriending schemes are available to the affected individuals (Bravata et al., 2023; Fakoya et al., 2021; Swartz, 2019; Thompson et al., 2016).

However, research suggests loneliness solutions must be further tailored to the user's needs so as to be more effective (Ypsilanti & Lazuras, 2022). In addition, past studies have identified issues with investigating loneliness as a phenomenon, such as the under-representation of adults aged 25-65, with most data coming from young (18-21) and older adults (65+); as well as the need to consider behavioural data in future research (Reedman-Flint et al., 2022). While research on the under-represented populations is emerging (Herrmann et al., 2023; Boden-Stuart et al., 2021; Bu et al., 2020) and methods are being implemented to offer tailored loneliness interventions (Morrish et al., 2023; Boekhout et al., 2021), the various angles of loneliness are often looked at in isolation from each other (e.g., Neves et al., 2025; Shen et al., 2025; Liu et al., 2022; Cacioppo & Cacioppo, 2018). 

Considering the above problems, this PhD project aims to deeply understand people's experiences of loneliness in relation to their sense of belonging in the social world from an interdisciplinary perspective. Bringing together data science, psychological and social sciences, it focuses on befriending schemes as a loneliness intervention tailored to the user's needs. Furthermore, this project seeks to gain insights into the factors associated with loneliness levels and the structure of an individual's social network in order to get a detailed picture of the context surrounding loneliness as an issue investigated from multiple angles, which has so far been done in isolation.

The proposed methodology will be of mixed nature. Using machine learning to identify key factors for successful befriending schemes, this PhD aims to provide guidelines for tailored solutions for this type of loneliness interventions. Furthermore, follow-up interviews with befriendees and befrienders who participate in befriending schemes will be used to gain a deeper understanding of the specific nature of individual schemes and to identify the recurring themes which might point towards a specific mechanism based on which this loneliness intervention succeeds. Finally, a data scientific approach will be extended to the identification of factors associated with the experienced, self-reported levels of loneliness, using neodemographic data cross-sectionally. 

The data will be analysed using a mixed-methods approach: combining predictive models; quantitative and qualitative language analysis methods, as well as social network analysis.

This PhD project will contribute to the interdisciplinary understanding of loneliness as well as befriending schemes as one of the most widespread loneliness interventions. The proposed approach will allow to design solutions for loneliness relief tailored to the user's needs, as understood from the factors surrounding loneliness; as well as the subjective perspectives of the individuals. Taking a step further, the project will provide guidelines for expanding befriending schemes as a loneliness intervention to make it a solution available to a wider group of service users.

The project runs in partnership with B:friend – a wellbeing charity offering support to lonely older adults. The work is supported by the Engineering and Physical Research Council [Grant number EP/S023305/1].