Amul's topical advertisements represent one of the world's longest-running contextual advertising which is successfully integrating contemporary socio-political, cultural, economic, and sporting events with a distinctive and recognizable visual identity. While previous research has extensively examined advertising effectiveness, consumer engagement, and brand communication, limited studies have quantitatively investigated the visual consistency of topical advertisements using Image Analytics. To address this research gap, the present study proposes a data-driven image analytics framework implemented in Orange Data Mining to analyze visual similarity patterns within a dataset of selected 60 Amul topical advertisements. Deep image embeddings were extracted from each advertisement to capture high-level semantic visual features, which were subsequently analyzed using Euclidean distance computation, distance matrices, hierarchical clustering, k-means clustering, t-distributed Stochastic Neighbour Embedding (t-SNE), and outlier detection. The proposed workflow enables an objective and scalable assessment of visual relationships among advertisements without relying on manual annotation, thereby providing a reproducible methodology for evaluating brand consistency through unsupervised machine learning. The experimental analysis demonstrates that although the advertisements represent diverse contemporary events and creative narratives, they exhibit a high degree of visual consistency through recurring illustration styles, typography, mascot representation, color composition, and layout design. The findings demonstrate the effectiveness of deep image embeddings and unsupervised learning for objectively assessing visual branding strategies. The proposed framework contributes to the growing intersection of computer vision, image analytics, and marketing analytics by providing a scalable and reproducible approach for evaluating visual consistency in advertising campaigns, with p