Abstract: Efficient and accurate assessment of elephant populations is critical for conservation efforts, particularly in Southern Africa where the Conservation Ecology Research Unit (CERU) at the Department of Zoology and Entomology operates. Current methods, such as Rapid Elephant Population Assessments (REPA), although effective, are labor-intensive and time-consuming. To address these challenges, this project explores data science techniques to automate population assessments. Utilizing high-resolution aerial images of elephant herds, this study develops a novel approach to predict elephant numbers and extract individual back lengths. Leveraging annotated datasets, two models, YOLOv5 and YOLOv8, were trained and evaluated. YOLOv8 demonstrated superior performance over YOLOv5, accurately identifying 107 out of 128 elephants and 77 out of 127 back lengths. Data preprocessing involved annotating images using a computer vision annotation tool, facilitating object detection training. However, challenges remain in automating the annotation process to enhance efficiency further. This project showcases the potential of data science in conservation efforts, offering a promising solution to streamline elephant population assessments, ultimately supporting informed management decisions and conservation strategies. Further advancements in automation and model refinement hold the key to optimizing this approach for broader application in wildlife conservation.