Data mining has many applications in computer sciences and engineering problems. This paper is focused on the application of data mining techniques to observe the effect of weather parameters on electric load in Pakistan. In power utilities, the factors affecting electric load can play a vital role in the design of different types of electric load forecasting models. In this paper, various data mining techniques have been applied on historical weather and electric load data and results have been obtained. The results have identified very interesting similarities in the electric load and certain weather parameters showing similar clusters over the years in the available data. The key weather parameters identified in the analysis are mainly temperature and relative humidity that have major impact on summer electric load and their effects were validated by an ANN (Artificial Neural Network) based STLF (Short Term Load Forecasting) model using PSO (Particle Swarm Optimization) learning rule. A considerable reduction in the forecasting error has been achieved confirming the efficacy of the data mining approach towards the impact of weather sensitivities on electric load.