Demographic data often misrepresents marginalized communities, whether it be through overgeneralizing experiences by aggregating data when it shouldn’t, omitting entire communities from the data altogether, or categorizing individuals as “other." Asian American Pacific Islander (AAPI) people are particularly at risk, being misrepresented in all kinds of data from cardiovascular risk data to education data. These issues are not only inherently harmful but also inflict more harm when problematic data is used to develop AI. In an investigation of data bias against AAPI people, I employed machine learning and data analysis on Census data: I analyzed the effects of disaggregating race data, compared various approaches to coding data on multiracial and multiethnic people, and studied the effects of including race data in AI.