From observed data, statistical inference infers the properties of the underlying probability distribution. For hypothesis testing, the t-test and some non-parametric alternatives are covered. Ways to infer confidence intervals and estimate goodness of fit are followed by the F-test (for test of variances) and the Mann-Kendall trend test. Bootstrap sampling and field significance are also covered.

Transformation of the Earth’s social and environmental systems is happening at an incredible pace. The global population has more than doubled over the last five decades, while food and water consumption has tripled and fossil-fuel use quadrupled. Attendant benefits such as longer lifespans and economic growth are increasingly joined by corresponding drawbacks, including mounting socioeconomic inequality, environmental degradation, and climate change. Over the past half-century, interregional differences in population growth rates, unprecedented urbanization, and international migration have led to profound shifts in the spatial distribution of the global population. Economic changes have been dramatic as well. The global per-capita gross domestic product doubled while economic disparities grew in many regions (Rosa et al. 2010).

For all the potential of big data and data science, scholars wanting to untangle human–environment interactions face many gaps in big data. Human-environment data handily exemplify many of the characteristics of big data. They have high volumes, orders of magnitudes more extensive than commonly used in most research fields, resulting from repeated observations over time and space (Jacobs 2009). These spatial and temporal data are often collected and analyzed across multiple scales. They exhibit high velocity, with data being collected and stored in or near real time from an extensive array of sensors at sea, on land, and in air and space, alongside data collected via social networks and internet sources. Human-environment data also exhibit incredible variety in the domains that they relate to and the structures and data models necessary to conduct research. They often represent complex social and biophysical entities and relationships that operate at multiple levels of an organization, over space, and through time. These data also push the boundaries of other characteristics, including value for answering specific questions and veracity in terms of accuracy and fitness for use.

Data science proponents sometimes contend that their approach augurs the “end of theory” because we are on the threshold of a new scientific paradigm. Big data is seen as a powerful “black box” into which users shovel large and messy data collections and, in return, get deep insights into any domain people care to investigate. The essence of black box research is that the inner workings of the computation can remain opaque to most users, and there may be little need to understand much about the process being modeled. Data science can pursue an inductive approach to knowledge creation that frees researchers from fieldwork or dealing with the messy business of deductive science that involves developing hypotheses and conceptual frameworks. The proposition of theory-free science has sparked great debate and attendant conversations on how science is practiced and moving beyond simple epistemological binaries.

Many of the challenges with human-environment data are exacerbated by critical methodological shortcomings. Many of the core tasks of data science, such as data manipulation, machine learning, or artificial intelligence, are made more difficult by the complex nature of human-environment data. Data science and cognate fields are the loci of exciting research in addressing these methodological challenges. Information science on data lifecycles is being adapted to the needs of data science, including research in the methods of metadata, ontologies, and data provenance. Computer science and related fields adapt approaches like parallelism and distributed computing to work with big human-environment data. There is also ongoing work in cloud computing and high-performance computing to address the needs of complex spatiotemporal data sets. The private and public sectors invest heavily in smart computing, embedded processing, and the Internet of Things (IoT). An extraordinary amount of effort is dedicated to sharing data and workflows to support reproducibility in science. Finally, data science has advanced how it handles data that capture spatial and temporal patterns and processes.

Data science plays an increasingly prominent role in examining and addressing human–environment dynamics. The growing role of data-intensive inquiry is apparent to many researchers, policymakers, and others. At the same time, however, there is mounting awareness of the interplay between the perils and promise of big data for human-environment systems in particular. Despite all the research and interest, there are few straightforward ways to reconcile the pros and cons of data science. This difficulty stems partly from the many unknowns about data science because its story is still being written. There are exciting developments in data science theory, conversations around democracy and decision-making, and promising developments in data science infrastructure.