Leveraging imagery data in evaluations: Applications of remote-sensing and streetscape imagery analysis

This paper discusses using imagery data in evaluations and the advantages and limitations of relevant methodologies.

Imagery data, such as remote-sensing images and digital photos, are becoming easier to work with thanks to better technology, including machine learning techniques and more powerful computing resources.

The paper suggests that image analysis can improve evaluations by clearly showing what's happening on the ground. A key advantage of image analysis is that it can be used to quantify changes in environment, infrastructure and human activity precisely over time and space.

It is noted that it's important to check image data against the reality on the ground, in cases where imagery data are being used as proxies for complex phenomena "for example, digital photos depicting the physical characteristics of houses can be used as a proxy for poverty levels". (Ziulu, 2024, p. 35)

The goal of this resource is to help evaluators use images effectively so they can focus their efforts on improving development activities.