Gentrification has become a critical issue in urban planning during the past half century. Ruth Glass first coined the term ‘gentrification’ in 1964 to describe the phenomenon that wealthier class replace lower-class in certain neighborhoods in cities. Often gentrification is portrayed as a tool of revitalization for those declining urban areas. It was not until 1979 when the negative consequences of gentrification, known as displacement, was fully understood. The negative socio-economic consequences of displacement including income inequalities and injustices were further investigated and evaluated in modern literature. Governments and non-profit organizations have taken efforts to actively intervene when displacement occurs through various policy responses including job training, apprenticeships, subsidized and transitional employment programs, equal opportunity laws, community reinvestment strategies and more. Besides all the efforts taken by the government after gentrification is detected, successful prediction of gentrification can help identify displacement at an early time and guide policy interventions before they become more costly and less effective. Traditional statistical methods have been applied to forecast gentrification when the negative outputs of gentrification were first noticed and came to social scientists’ attention. In 1980s and 1990s, integrating data from various sources, neighborhood early warning systems were established in big cities, including Chicago Neighborhood Early Warning System, Neighborhood Knowledge Los Angeles, the Philadelphia Neighborhood Information System and so on. Modeling methods take advantage of these systems to provide predictive analysis mainly for municipal agencies. For instance, logistic regression is used to “analyze data from the Philadelphia Neighborhood Information System in order to determine which properties were most likely to become imminently dangerous”. Multivariate regression is also a simple while commonly used method to identify key factors that make a neighborhood to gentrify. These traditional techniques often perform poorly as common measurements of gentrification such as property value cannot be modeled linearly with other features using traditional methods. Some models also fail due to high correlations among variables or autocorrelations in residuals. Therefore, it is necessary to explore innovative predicting algorithms and evaluate their performances. Some studies have sought to employ new algorithms and have found out that techniques like machine learning usually outperform those traditional ones. This project explores how well a range of machine learning models perform on predicting neighborhood gentrification with socio-economic data rather than applying traditional statistical models.