Reducing return temperatures in district heating (DH) networks is crucial for improving overall efficiency, minimising thermal losses, and enabling the integration of renewable energy sources. Achieving this requires a detailed understanding of DH substations performance, as faults or inefficiencies in substations are a common cause of elevated return temperatures. This case study introduces a data-driven approach for fault detection and diagnosis (FDD) in DH substations, aiming to identify such inefficiencies and support targeted performance improvements. The methodology is implemented in a Python-based tool that analyses and visualises smart meter data collected from the primary side of substations. The tool offers geospatial mapping, performance indicators (e.g., return temperature, hydraulic capacity utilization), and metadata filtering (e.g., building type). Using this approach, four recurring fault patterns are identified. To support fault handling prioritisation, the tool segments inefficient substations through scatterplot visualisations. The proposed approach offers utility companies a means to enhance system reliability, reduce emissions, and improve operational performance by systematically identifying and addressing performance issues.