This study extends Media Richness Theory (MRT) by examining how emojis function as nonverbal cues in digital sport communication. This was accomplished through a mixed-method approach combining machine learning and quantitative content analysis of 13,642 X (formerly Twitter) posts from professional sport teams. The study aimed to understand how emojis function within sport organizations' digital messaging and to identify contextual factors that influence emoji interpretation in sport digital communication. Our findings indicate that emojis serve dual functions: (i) as replacements for textual content and (ii) as supplements that reinforce emotional resonance with the classification models, achieving higher accuracy for singleemoji (80.3% F1-score) versus multi-emoji messages (72.2%). Contextual elements, including immediate textual surroundings and broader sport situations, significantly shape emoji interpretation and effectiveness. Results show that strategic emoji usage patterns vary across game situations, team performance contexts, and marketing-related announcements. This research offers theoretical contributions by extending MRT to account for nonverbal digital cues while providing foundational understanding to inform sport marketing strategies. The findings establish patterns of emoji functionality that future research can build upon to directly measure consumer response and marketing outcomes across diverse sport contexts.