Epsilon Archive for Student Projects

Abstract

This thesis is about establishing sustainable housing gardens in urban residential areas to promote
well-being among residents. It is also about how people´s values and needs in life is shaping their
perception and how it relates to use of green space. The overall purpose has been to investigate how
people´s values and basic human needs can be used for impacting the design of urban green space.
The outcome of this thesis work aims to provide relevant insights useful in further research about
the development of new green areas for multi-family homes in larger cities. The aim is to cover
aspects on how green space in cities can be designed using the Perceived Sensory Dimensions (PSD)
model (Grahn & Stoltz, 2021) when shaping green environments, with people in mind. The research
aim has also been to explore how people´s values affect their perception of different green
environments surrounding urban residential buildings and how Barrett’s Values Model can be
related to the PSD framework. By exploring the use of new AI tools as part of this study, the
ambition was to provide inspiration for further research on how landscape architects can utilize AI
in relation to existing PSD research to promote health and wellbeing among residents in dense cities.

The first part of this master’s thesis includes a summary of established research and frameworks to
highlight valuable insights about how green space can impact the overall wellbeing of people. The
intention by outlining some established environmental perception frameworks is to show how it can
be used in future urban planning and property development for residential housing in cities. It also
to includes a summary of relevant research on values and how it relates to psychological needs of
humans in the context of developing sustainable cities. The second part has been to use one of these
research frameworks – Perceived Sensory Dimensions – developed by Grahn and Stoltz (2021), by
testing AI-generation of photographs to show examples of green environments in urban residential
housing areas. The last part was to test these AI generated photos by conducting a quantitative pilot
survey to identify if and how people’s perception differs depending on their values in relation to
Barrett’s model and how it related to perception and the use of green space.

The conclusions were that values can be used in alignment the environmental frameworks to further
explore human needs and what is important for various user groups in different life stages
concerning both restorative and stimulating environments. From the findings it seems that Barrett's
value model and the PSD model can be related as higher values seem to relate to stimulating PSDs
Diverse, Open, Cultural and Social, whereas lower values seem to connect more to restorative PSDs
Serene, Sheltered, Natural and Cohesive. Correlations were found between different value groups
and how they perceived different photos of PSD environments, which indicate that personal values
can affect environmental preference. The PSD research definitions could be for prompting in the
Midjourney AI bot to generate photos of green environments covering nearly all PDS environmental
qualities, except for Cohesive. Most environments could be recognized in the descriptions as well
as evoked feelings by the survey respondents in relation to the Perceived Sensory Dimensions.
Serene and Diverse environment scored the highest on preference. By using environmental research
frameworks and integrate it with generative AI tools, there will be future opportunities for urban
planners, landscape architects and property developers to enhance green spaces to promote health
and well-being.