The article introduces the R package `np`, which implements a variety of nonparametric and semiparametric kernel-based estimators. The package is designed to handle mixed data types, including continuous, discrete, and categorical variables, and provides data-driven methods for bandwidth selection. Key features include unconditional and conditional density estimation, mean and quantile regression, model specification tests, and significance tests. The article highlights the package's flexibility and ease of use through various empirical applications, comparing parametric and nonparametric methods in regression, binary outcome models, and conditional probability estimation. It also discusses the importance of bandwidth selection and provides guidance on using the package's functions, including creating custom estimators and handling mixed data types. The article emphasizes the practical benefits of nonparametric methods, such as better capturing underlying structures in data and improving predictive accuracy.The article introduces the R package `np`, which implements a variety of nonparametric and semiparametric kernel-based estimators. The package is designed to handle mixed data types, including continuous, discrete, and categorical variables, and provides data-driven methods for bandwidth selection. Key features include unconditional and conditional density estimation, mean and quantile regression, model specification tests, and significance tests. The article highlights the package's flexibility and ease of use through various empirical applications, comparing parametric and nonparametric methods in regression, binary outcome models, and conditional probability estimation. It also discusses the importance of bandwidth selection and provides guidance on using the package's functions, including creating custom estimators and handling mixed data types. The article emphasizes the practical benefits of nonparametric methods, such as better capturing underlying structures in data and improving predictive accuracy.