Big Data presents new opportunities and challenges for data scientists. While it enables the discovery of subtle population patterns and heterogeneities, it also introduces computational and statistical challenges such as scalability, storage, noise accumulation, spurious correlation, incidental endogeneity, and measurement errors. These challenges require new statistical and computational paradigms. This paper discusses the salient features of Big Data and their impact on statistical and computational methods, as well as computing architectures. It emphasizes the viability of the sparsest solution in high-confidence sets and highlights that exogenous assumptions in most statistical methods for Big Data cannot be validated due to incidental endogeneity, which can lead to wrong statistical inferences and scientific conclusions. The paper also provides new perspectives on Big Data analysis and computation, including the development of new statistical procedures and computational infrastructure. It discusses the challenges of analyzing Big Data, such as heterogeneity, noise accumulation, spurious correlations, and incidental endogeneity, and the need for new statistical thinking and computational methods to address these challenges. The paper also covers the impact of Big Data on statistical thinking, computational methods, and computing infrastructure, as well as the challenges of analyzing Big Data in genomics, neuroscience, economics and finance, and other applications. It concludes with the importance of understanding the unique features of Big Data and the need for new statistical and computational methods to address these challenges.Big Data presents new opportunities and challenges for data scientists. While it enables the discovery of subtle population patterns and heterogeneities, it also introduces computational and statistical challenges such as scalability, storage, noise accumulation, spurious correlation, incidental endogeneity, and measurement errors. These challenges require new statistical and computational paradigms. This paper discusses the salient features of Big Data and their impact on statistical and computational methods, as well as computing architectures. It emphasizes the viability of the sparsest solution in high-confidence sets and highlights that exogenous assumptions in most statistical methods for Big Data cannot be validated due to incidental endogeneity, which can lead to wrong statistical inferences and scientific conclusions. The paper also provides new perspectives on Big Data analysis and computation, including the development of new statistical procedures and computational infrastructure. It discusses the challenges of analyzing Big Data, such as heterogeneity, noise accumulation, spurious correlations, and incidental endogeneity, and the need for new statistical thinking and computational methods to address these challenges. The paper also covers the impact of Big Data on statistical thinking, computational methods, and computing infrastructure, as well as the challenges of analyzing Big Data in genomics, neuroscience, economics and finance, and other applications. It concludes with the importance of understanding the unique features of Big Data and the need for new statistical and computational methods to address these challenges.