The escalating demand for mobile broadband poses significant challenges for spectrum management, particularly in developing countries where detailed traffic data are often inaccessible due to concerns over privacy and confidentiality. To address this issue, it is essential to adopt a data-driven and analytical approach to spectrum planning. This study proposes a Quality of Experience (QoE)-driven three-stage framework that leverages average downlink throughput as a practical proxy for forecasting future mobile broadband spectrum needs. The framework integrates time-series forecasting using statistical, machine learning, and deep learning models; spectrum gap analysis, which translates the forecasted QoE into a quantitative demand versus current network supply; and data-driven spectrum policy recommendations. Using 60 months of QoE data, which were interpolated into 257 weekly samples from four mobile network operators (MNOs) in Jakarta, Indonesia, the framework identified significant spectrum deficits for three MNOs, requiring at least 15 MHz in the 700 MHz band and 20 MHz in the 2.6 GHz band to meet the projected demand. The proposed framework offers a transparent, replicable, and data-driven methodology that enables regulators and MNOs to make evidence-based spectrum allocation decisions, providing a practical solution to the persistent problem of data inaccessibility and supporting more proactive and equitable national spectrum planning. © 2013 IEEE.