This study analyzes the collaboration network of jazz musicians using a database of jazz recordings. The network is constructed at two levels: one focusing on individual musicians and another on bands. The individual network connects musicians who have played in the same band, while the band network connects bands that share at least one musician. The analysis reveals that both networks capture essential aspects of social interactions among jazz musicians. The study finds correlations between community structure, recording locations, and racial segregation. The degree distribution of the musician network shows a power law tail, while the band network exhibits a stretched exponential distribution. The musician network is highly clustered, with a small-world property, and shows assortative mixing in degrees, indicating that high-degree nodes are more likely to connect to other high-degree nodes. However, the band network does not show such correlations. The community structure analysis reveals racial segregation between black and white musicians, with the majority of bands being exclusively one race. The community structure also correlates with recording locations, with New York and Chicago being dominant recording locations. The community size distribution of the jazz network is found to be similar to that of an email network, showing a power law decay with the same exponent up to a certain size. The study highlights the importance of community structure in understanding social networks and suggests that future research could explore the fundamental laws governing the formation of community structures in social interactions. The results show that both the musician and band networks capture essential ingredients of the collaboration network of jazz musicians.This study analyzes the collaboration network of jazz musicians using a database of jazz recordings. The network is constructed at two levels: one focusing on individual musicians and another on bands. The individual network connects musicians who have played in the same band, while the band network connects bands that share at least one musician. The analysis reveals that both networks capture essential aspects of social interactions among jazz musicians. The study finds correlations between community structure, recording locations, and racial segregation. The degree distribution of the musician network shows a power law tail, while the band network exhibits a stretched exponential distribution. The musician network is highly clustered, with a small-world property, and shows assortative mixing in degrees, indicating that high-degree nodes are more likely to connect to other high-degree nodes. However, the band network does not show such correlations. The community structure analysis reveals racial segregation between black and white musicians, with the majority of bands being exclusively one race. The community structure also correlates with recording locations, with New York and Chicago being dominant recording locations. The community size distribution of the jazz network is found to be similar to that of an email network, showing a power law decay with the same exponent up to a certain size. The study highlights the importance of community structure in understanding social networks and suggests that future research could explore the fundamental laws governing the formation of community structures in social interactions. The results show that both the musician and band networks capture essential ingredients of the collaboration network of jazz musicians.