Dynamic Monitoring of Google Play Store Ranking Changes: Data Mining of Important Features

This study investigates the principal factors influencing keyword rankings of gaming applications on Google Play, thereby unveiling the operational dynamics of the Google Play recommendation system and elucidating the determinants of top ten game app rankings. The research commenced by categorizing seven keyword groups and selecting 24 keywords for bi-hourly data scraping over a two-week period, yielding 174 crawls that captured fluctuations in rankings. Through the deployment of crawler scripts and logical processing, amassed 104,849 data entries from 865 applications. Strategic model selection and feature optimization were pivotal, notably through the integration of textual data from titles and descriptions and the computation of time-series attributes, which escalated the model’s precision from a baseline of 0.9431 to 0.977 within 5 improvements. Comprehensive data mining uncovered that high download volumes, stable high review rates, and upload video introductions significantly contribute to a top-ten ranking. Additionally, refining the management strategy of top reviews could bolstered user engagement and satisfaction. Importantly, while existing scholarly work on ASO strategies for Google Play typically relies on static data, this study bridges a critical research gap by dynamically monitoring app changes to forecast keyword rankings.