The ongoing transformation of the Nordic power system, driven by the increasing share of renewable generation and the implementation of Flow-Based Market Coupling (FBMC) in October 2024, has introduced new challenges for electricity price forecasting. To improve short- term price models, market participants require an accurate representation of network sensitivities, which depend on the estimation of Power Transfer Distribution Factors (PTDFs). A key input in these calculations is Generation Shift Keys (GSKs), which define whether a node responds to a price change and participates in a change of a bidding zone’s net position. The choice of a GSK strategy directly influences how PTDFs are represented in market models, thus impacting the capacity allocation and market results. This thesis, conducted at Vattenfall AB, investigates how data-driven GSKs based on actual generation data affect the accuracy of zonal PTDF calculations within the Nordic day-ahead electricity market. Three time-resolution levels – monthly, weekly, and daily – were evaluated and compared against a simplified flat GSK reference. Each GSK was computed using hourly production data and complemented with internal capacity documentation. The PTDFs were calculated using Vattenfall’s grid model based on PyPSA which is an open- source framework for simulating and optimizing electrical networks, and performance was assessed using the Root Mean Square Error (RMSE) between modeled and official PTDFs from the Joint Allocation Office (JAO). The results show that all three data-driven approaches outperform the flat reference by improving PTDF accuracy across most bidding zones. Among them, the monthly GSKs achieve the best overall balance between accuracy and numerical stability, while the weekly and daily GSKs capture short-term variations more dynamically but at the cost of increased volatility. These findings suggest that even relatively simple, data-based GSK formulations can enhance the reliability of power flow representation in flow-based models. Improved PTDF estimation can in turn contribute to more accurate flow-based capacity calculations and, ultimately, to more reliable short-term electricity price forecasts.