Water is considered as a vital limited resource worldwide. Due to its geographical location, Iran suffers from a semi-arid and arid climate. In recent years, due to excessive use of water resources and long-lasting drought, the country has faced severe water scarcity. In order to deal with such a problem, the country needs to have proper water resource management strategy and practice. One of the most critical issues in this regard is related to the monitoring and management of urban water consumption. Exploring the pattern of urban water consumption and the relationships between geographic and demographic parameters and water usage is an essential requirement for effective management of water resources. In this study, association rule mining has been used to analyze the data of public water consumption in the city of Esfahan. Association rules mining have been used to discover the connections between geographic and demographic parameters including the number of family members, the number of apartment units, residential building types, areal coverage of the house and green spaces, distance from river and population centers, spatial location, distance from main roads, population density and percentage of young population with water consumption patterns. A version of the apriori algorithm called Liu, with suitable computational characteristics to process large amounts of data, has been used for association rule mining. This algorithm, using classification methods, provides the possibility to extract a broader range of association rules from the data. The output of the algorithm is a set of rules that can be subjected to study further using statistical methods. Each of the extracted rules that satisfy minimum support equal 30 and minimum confidence equal 60 reflects the relationship of each spatial and demographic parameters with the consumption water. Then, the obtained rules have been evaluated, and water consumption hot spots were extracted. The evaluation procedure consists of two parts. The first part examines the spatial pattern of household water consumption distribution. The second part investigates the distribution pattern of household water consumption by using Moran’s spatial autocorrelation analysis and identifies hot spots of consumption. Getis Ord's Gi test and hot spot analysis have been used to determine hot spots. Results show that some of the geographic and demographic parameters are associated with reducing consumption and some with increasing consumption. Distance from the main road, areal coverage of the residential and green spaces have a direct relationship with household water consumption. On the other hand, the number of residential units, population density, distance from the river, X and Y values, and percentage of the young population are inversely related to household water consumption. Also, if the distance from the city center increases, household water consumption decreases. Moreover, by moving to the north and east, the consumption of each unit reduces, and the southern areas of the city by minimum distance from the river have the highest water consumption and specified as hot spots. In these neighborhoods, building types are villa and have a lower population density with larger areas of green spaces and yards.