The sexual network topology of individuals and their communities may affect the spread of infectious disease both within and across specific populations. To better understand this process, we collected ego-centric data on the personal sexual networks of 175 young men who have sex with men. Respondents were asked to nominate sexual contacts and the sexual contacts of their nominees (n=602). Among the 741 men in the sample, 360 were Black, 164 Latino, 156 White, and 61 other race/ethnicity. Researchers often use ego-centric network data to investigate hard-to-reach populations, especially individuals living with or at risk for HIV. Ego-centric data, however, are typically incomplete, which makes studying disease spread difficult. For this reason, a semi-supervised entity-resolution scheme was designed in order to match data for unique individuals. Entity resolution yielded a reconstructed network with 628 observed sexual ties. To model the network we assume an exponential random graph models (ERGM). To account for not having complete information on some of the alter-alter ties we fit the model using a Bayesian data augmentation algorithm. The algorithm also provides us with a distribution of networks with imputed ties, allowing for further investigation of structures that are crucial for studying disease spread. We modeled race/ethnicity and age on the individual level; median income and distance were calculated on the neighborhood level. We can conclude that one source of racial discrepancy might be that Black participants have more sex partners than White participants, and that this is further amplified by strong racial homophily. There is strong evidence of clustering by neighborhood with respect to geographical distance. Triadic closure is less prominent within the Black and Latino community which provides indirect evidence for endogenous network processes being responsible for higher HIV spread within these communities.