Predicting geographical sites using microbiome data has been envisioned by scientists for decades but has been hampered by the lack of data and ignorance about the forces governing biodiversity. Over the past decade, next-generation sequencing data generated by large microbiome projects showed that taxa exhibit non-uniform patchy geographical distribution and that taxa can be classified into cities with high accuracy. We develop tools that identify and map microbiome throughout the world. This requires comparing our reference data to large microbiome datasets, which requires high computing time and storage.