Data Management Best Practices

Abstract

Data management best practices Properly managing data is of the upmost importance to ensure data quality and integrity, while also improving data access and discovery. To be successful, proper data management must be incorporated in all stages of data processing including collection, analysis, and long-term storage. The Washington University in St. Louis Libraries, has outlined step by step procedures for researchers to undertake to improve their data management techniques. Recommendations: 1. Conduct a data asset inventory. a. Assess and inventory data collected b. Create a standard data intake form to document details on data collection entities and methods for current and future data acquisitions 2. Adopt best practices for managing data a. Create an intuitive folder structure and hierarchy that reflects the project goals b. Adopt and apply consistent file naming conventions for all files and folders c. Actively manage digital objects d. Establish a robust backup strategy (preferably 3 copies-2 locally-1 offsite) 1. Inventory data An incremental, stepped approach to undertaking a data inventory is described below that can be applied flexibly according to context and needs. 1. Plan and define the purpose and scope of the inventory (determine which projects, whether or not to include off-site collaborators, what variables to include in the inventory, etc.). 2. Identify: a. what data assets exist, b. the values of the variables, c. where management could be improved, d. and any classifications of the data. Free, open source software can help automate the file inventory and identification process. FITS: http://projects.iq.harvard.edu/fits BitCurator: http://wiki.bitcurator.net/?title=Software Record the outputs of your inventory on a spreadsheet. Variables to track may include: • Dataset location (folder path) • Dataset File Name • Date of data collection • Project Name • Data Type (interview recording, transcript, etc.) • Data status (active, raw, analyzed, closed) • Collection Method (instrument, survey, interview, etc.) • Name of data collector