This paper presents a method for detecting community structure in directed networks by incorporating edge direction information into the modularity function. Traditional methods for community detection in undirected networks often ignore edge direction, which can lead to loss of important structural information. The authors generalize the modularity function to account for edge directions, allowing for more accurate community detection. The modularity function is defined as the difference between the fraction of edges within communities and the expected fraction of such edges under a random model. For directed networks, the expected fraction is calculated using in- and out-degrees of nodes. The authors propose a directed version of the modularity matrix and show that maximizing this matrix leads to better community detection. The method is based on spectral analysis, where the eigenvectors of the modularity matrix are used to determine community assignments. The algorithm is tested on various networks, including real-world examples such as the web and a network of technical terms. The results show that the directed method outperforms undirected methods in detecting community structure, especially in cases where edge direction is important. The algorithm is applied to both real and simulated networks, demonstrating its effectiveness in recovering known community structures and identifying new ones. The method is computationally efficient and can be applied to both small and large networks. The authors conclude that the directed modularity method is a valuable tool for analyzing directed networks, as it makes explicit use of edge direction information, which is often discarded by other methods.This paper presents a method for detecting community structure in directed networks by incorporating edge direction information into the modularity function. Traditional methods for community detection in undirected networks often ignore edge direction, which can lead to loss of important structural information. The authors generalize the modularity function to account for edge directions, allowing for more accurate community detection. The modularity function is defined as the difference between the fraction of edges within communities and the expected fraction of such edges under a random model. For directed networks, the expected fraction is calculated using in- and out-degrees of nodes. The authors propose a directed version of the modularity matrix and show that maximizing this matrix leads to better community detection. The method is based on spectral analysis, where the eigenvectors of the modularity matrix are used to determine community assignments. The algorithm is tested on various networks, including real-world examples such as the web and a network of technical terms. The results show that the directed method outperforms undirected methods in detecting community structure, especially in cases where edge direction is important. The algorithm is applied to both real and simulated networks, demonstrating its effectiveness in recovering known community structures and identifying new ones. The method is computationally efficient and can be applied to both small and large networks. The authors conclude that the directed modularity method is a valuable tool for analyzing directed networks, as it makes explicit use of edge direction information, which is often discarded by other methods.