Design and Usability Evaluation of a Weather-Aware Mobile Movie Recommendation System Using a Human-Centered AI Approach

Abstract

The global digital entertainment industry demands personalized, context-aware recommendation systems that respond to users' immediate environmental circumstances. NNG Cinema is a weather-aware mobile movie recommendation system developed using Flutter 3.16 and Dart 3.2, integrating real-time meteorological data from the OpenWeatherMap API with movie discovery capabilities from The Movie Database (TMDB) API. Grounded in Context-Aware Recommender System (CARS) theory, mood management theory, and human-centered AI design principles, the system automatically maps seven meteorological condition categories to curated genre clusters derived from environmental psychology literature, requiring no additional user input. This study employed a six-phase design-and-build methodology. Twenty-four black-box functional test cases were executed across six categories in three cycles, and 30 purposively sampled participants completed the System Usability Scale (SUS) following a single ten-minute session of free system interaction. All 24 test cases achieved Pass status across all three cycles (100% pass rate). SUS evaluation produced a mean score of 80.33 (SD = 7.84), placing NNG Cinema in the Good-to-Excellent usability range and 12.33 points above the established population median. Findings indicate that fully automated environmental context integration is technically feasible and elicits a positive usability perception as a basis for mobile movie recommendation. These results should be read as evidence of design feasibility and perceived usability only: recommendation accuracy, perceived relevance, user satisfaction, and technology acceptance in the sense of TAM or UTAUT were not measured, and no comparative baseline was evaluated.