Publicly-available repositories of biological data are incredibly valuable resources with the potential to improve our understanding of human health and disease. Unfortunately, these data are typically neither easily accessible nor well-annotated, limiting their utility in addressing biological questions. To overcome this, we have developed two resources, recount2 and phenopredict. recount2 increases accessibility by aggregating and processing data from 70,000 human RNA-sequencing samples on a single analytic pipeline. phenopredict improves annotation using in-silico phenotyping, a novel method that provides critical sample information across these 70,000 samples. By removing barriers to access and improving annotation of publicly-available data, we provide a valuable and user-friendly resource from which biological insights can be drawn.