This paper presents a practical method for estimating the hourly demand profiles of different loads at a feeder with basis on integrating low resolution maximum demand records, the variability of the characteristic demand and the maximum monthly curves with hourly resolution. The artificial samples of power demand are generated employing a uniform probability distribution with a range between the recorded maximum and the estimated minimum demand. The method was validated with a study case based on the data from a feeder with 2304 clients provided by a network operator in Colombia, the data did not present a fixed sample frequency. The proposed method was applied to generate load profile estimations for the IEEE 13-bus test feeder at each load point. The results show that the method is useful to generate load profiles with hourly resolution, allowing to solve the lack of recorded data at the load points of a feeder.