China
Africa
Explore chapters and articles related to this topic
Assessing Provenance and Bias in Big Data
Published in Diane P. Michelfelder, Neelke Doorn, The Routledge Handbook of the Philosophy of Engineering, 2020
Brent Mittelstadt, Jan Kwakkel
Presently, a variety of data provenance management and process provenance (i.e. workflow) management systems exist, such as VisTrails and Kepler. For a comprehensive overview, see Pérez et al. (2018). These systems all involve the automated capturing of provenance-relevant details including data processing steps, execution information, and any additional user specified annotation that provide a motivation for specific choices or an interpretation of intermediate results. These systems also offer some way of representing provenance information. The representational model is typically domain-specific and layered. A layered representation enables presenting provenance at different levels of abstraction. For example, one layer might capture the overall workflow, while another layer captures individual traces of the execution of individual steps of this workflow. Storing and querying is the third component of these management systems. Here different researchers are experimenting with domain-specific querying approaches, as well as querying mechanisms dictated by the storage format (e.g. SQL). That is, can a query of the provenance data be expressed in a language familiar to the user, or must it be expressed in a language that is related to the database in which the provenance data is being stored? The advantage of the former is that it enables the reuse of domain language and abstracts the storage system away from the user.
Urban and building multiscale co-simulation: case study implementations on two university campuses
Published in Journal of Building Performance Simulation, 2018
Clayton Miller, Daren Thomas, Jérôme Kämpf, Arno Schlueter
The coupling, simulation and co-simulation process is completely managed within a program called VisTrails (Freire et al. 2014). The implementation of this type of workflow is outlined in previous work focused on automating DPV using the Kepler platform (Thomas and Schlueter 2012). This process reduces the effort expended by a designer down to pressing a single button. This functionality allows an iterative design informed by readily available simulation results. An automated workflow ties the various steps together and maintains the effortless iterative design process. VisTrails enables the coupling of various workflow subprocesses script initializations, executions of the engines, and the compilation of the outputs. This tool empowers the coupling of multiple executable files and their connecting scripts in graphical diagrams that enhance reproducibility and process automation (Freire et al. 2014). The VisTrails workflow diagram for the first case study in this paper is seen in Figure 5.