Data Science Resource Core

点击翻译按钮获取中文摘要

The Data Science Core, based at Stanford University, is the component through which the overall management and sharing of experimental and simulation data will be conducted. The Data Science Core will implement the Data Science Plan by developing an integrated software framework for data storage, analysis, and simulation, and by closely collaborating with all team members to ensure that their research needs are met in a timely manner. This effort will require a full-time data scientist that will be responsible for implementing and overseeing the data management framework and ensuring its effective use among team members. The data management framework, detailed in the data science plan, is based on a common description and storage format for experimental datasets produced by the proposed experiments in Projects 1-4, and will include adapted versions of the existing analysis software or new analysis tools that support the common format, software tools to efficiently extract and represent cellular and network properties from experimental time series, and a pipeline to use data stored in these formats to constrain our large-scale neuronal network models in Project 5. The principal data scientist will be responsible for collaboration with external organizations that develop scientific data formats, such as Neurodata Without Borders and the HDF Group, in order to ensure that best technical practices are followed in development of the storage format and support software, and for effective dissemination to the broader scientific community of all software and data generated by the projects via a resource such as Collaborative Research in Computational Neuroscience (CRCNS). The Data Science Core will play an important part in achieving the overall goals of the research projects by ensuring consistent use of analysis methods, facilitating data sharing among team members, allowing direct comparison of the outcomes of experiments performed in different labs under different conditions, and will accelerate the development of open-source software tools to help increase reproducibility across research teams.