Energy scenario analyses are able to provide insights into the future and possible strategies for coping with challenges such as the integration of renewable energy sources. The models used for analyzing and developing future energy systems must be simplified, e.g., due to computational constraints. Therefore, grid-related effects and regional differences are often ignored. We tackle this issue by presenting a new methodology for aggregating spatially highly resolved transmission grid information for energy system models. In particular, such approaches are required in studies that evaluate the demand for spatially balancing power generation and consumption in future energy systems. Electricity transmission between regions is crucial, especially for scenarios that rely on high shares of renewable energy sources. The presented methodology estimates transmission line congestions by evaluating the nodal price differences and then applies a spectral clustering on these particular link attributes. The objective of the proposed approach is to derive aggregated model instances that preserve information regarding electricity transmission bottlenecks. The resulting models are evaluated against observables such as the annual amount of redispatched power generation. For a selection of defined performance indicators, we find a significantly higher accuracy compared to the commonly used, spatially aggregated models applied in the field of energy scenario analysis.