The Social Capital Atlas (Atlas) is a novel index from Chetty and colleagues that allows for the scientific study of social capital—a key driver of social inequalities that has nevertheless been understudied for its connection to social and health outcomes. The Atlas was defined at the ZIP Code Tabulation Area (ZCTA) level, a granular geospatial category derived from postal codes. However, ZCTAs have known limitations, and far fewer social, economic, and health indicators of interest are available at the ZCTA level compared to the census tract level. We therefore reconstruct the Atlas at the census tract level to provide a novel dataset that can be used to advance the scientific study of social capital. To accomplish this, we used block-level population weights to approximate tract-level estimates of economic connectedness (EC), support ratio (SR), and volunteering rate (VR), and then assessed their distributions and associations with upward mobility. Reconstructed distributions were highly similar to the original analysis, albeit with slightly lower means. EC remained strongly associated with upward mobility at both the ZCTA and census tract levels, including after adjustment for socioeconomic covariates. SR and VR showed weaker and less consistent associations. In sum, we present a reconstructed tract-level alternative to the public-use ZCTA-level Atlas that may be useful for public health, sociological, and policy research. However, these measures should be interpreted as geographic approximations rather than direct tract-level social network estimates.