Despite urbanization benefiting modern society and the people living in the urban city, the limited public resources, especially parking resources, remain a problem. Parking pricing acts as a tool to adjust the available resources. A logical question is: How to use parking pricing to maximize parking resource utilization while optimizing the parking revenue for parking management? In this MSc thesis, I propose an architecture that utilizes available public resources while optimizing revenue with predefined restrictions, especially in the parking management field. More specifically, I first (a) design a data-driven time series based prediction model, and then (b) design a reinforcement learning based dynamic pricing model to incorporate price restrictions. Moreover, I also (c) come up with metrics to evaluate the dynamic pricing model, as well as (d) implement and evaluate the proposed models with real parking data. Evaluation results show the effectiveness and practicality of my predictive analytics architecture for dynamic pricing for parking applications.