Oyinkansolaoluwa Ifidon-Ola

Oyinkansolaoluwa Ifidon-Ola

Biography

Oyinkan Ifidon-Ola is a senior at the University of Michigan pursuing a dual degree in Art & Design and Information Analysis. With a background in industrial design and a research-driven approach, she is passionate about leveraging data native design to create equitable, community-centered solutions. Her work bridges qualitative and quantitative methods, combining interviews, metrics, and spatial analysis to inform everything from assistive tools to urban planning strategies. She believes that design should adapt to communities, not impose on them. Oyinkan’s recent projects span UX research, data visualization, digital accessibility, product design, and creative programming. Her practice centers marginalized voices, informed by lived experience and a commitment to informal education. Committed to making complex systems more accessible, Oyinkan hopes to bring her interdisciplinary skills to a graduate program in architecture and planning, where she can help reshape the built environment through data native, community-driven design.

Abstract

Scaling LiDAR Analysis in Informal Settlements: An Automated Machine Learning Based Approach

Urban planners and policymakers require detailed spatial data on informal settlements to enhance living conditions and deliver essential services to rapidly growing populations. However, traditional mapping methods are often inadequate in these environments due to high building density, narrow footpaths, and uneven terrain. While Light Detection and Ranging(LiDAR) technology offers solutions through three-dimensional point cloud generation, current techniques rely heavily on manual processing, which requires substantial human oversight and makes large-scale analysis slow. Additionally, existing automated classification systems mainly focus on well-structured urban areas and lack frameworks tailored to the unique features of informal settlements. Here, we introduce an automated machine learning based approach for LiDAR point cloud classification in informal settlements. Using data from the Vidigal favela in Rio de Janeiro, Brazil, we implement a manual segmentation technique that classifies the point cloud into prevalent parts of informal settlements. We then use these segmentations to train machine learning classifiers. The eventual automated workflow is expected to reduce processing time from weeks to hours while maintaining classification quality comparable to that of manual methods, thereby enabling scalable analysis of informal settlements for urban planning purposes.