Repository logo

Contextualising energy justice in low-income built environment: Towards data-driven policy interventions for addressing distributive injustices in slum rehabilitation housing of the Global South



Change log



Around a billion people live in slums today globally, and rehabilitating them to formal housing is a significant challenge. Slum rehabilitation housing is a policy effort to solve this crisis and alleviate urban poverty. However, the question of whether slum rehabilitation programmes are accomplishing more good than harm or whether they are creating a whole host of new problems remains unexplored in the literature. This thesis investigates the effect of slum rehabilitation on household energy demand in Brazil, India and Nigeria through the lens of distributive energy justice. Furthermore, this thesis makes methodological innovation to aid in just policy design by improving the objectivity of including local and contextual knowledge on how poor households live and use energy. Doing so makes novel theoretical and methodological contributions: a theoretical contribution to temporality and spatial energy justice studies on how to offer cross-sectional depictions of energy demand within the slum rehabilitation housing, which was evaluated through structural equation modelling, and a methodological contribution in developing a deep-narrative analysis framework using natural language processing and machine learning-based Latent Dirichlet Allocation algorithm to capture the grounded narratives of distributive injustices objectively.

This research highlighted the significance of contextualisation in planning for energy justice in slum communities and the role of digital tools like natural language processing in objectively integrating grounded narratives in just policy design. The contextualisation was done through zoom-in and zoom-out of the grounded narratives enabled through the multi-method approach. Zooming-out view of distributed injustices in the study areas of Mumbai (India), Rio de Janeiro (Brazil) and Abuja (Nigeria) revealed inefficiencies in the administration of electricity distribution companies, lumped billing periods and lack of people-centric built environment design considerations. Similarly, zooming-in the case studies revealed that the poor design of the slum rehabilitation-built environment influenced the increase in energy intensity in the Mumbai case, leading to energy poverty. Whereas created distinct poverty traps in the Brazilian and Nigerian cases through frequent power cuts, high cost of appliance repair, and poor housing design. Finally, policy implications were drawn as per the policy actors across municipal, state and national levels that suggested leveraging digital tools like the deep-narrative analysis and the heavy penetration of Information and Communication Technology devices in such low-income communities. Such tools can improve accountability in decision-making and improve the representation of the occupants through their narratives of injustices associated with living in such communities. Thus, this thesis uniquely forwarded a data-driven pathway for integrating local collective intelligence in just policy design.





Sunikka-Blank, Minna


Energy Policy, Poverty, Slum Rehabilitation Housing, Natural Language Processing, Machine Learning, Energy Justice, Collective Intelligence


Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge
Bill and Melinda Gates Foundation through the Gates Cambridge Scholarship under the Grant Number OPP1144.