Conducting research involves working with data and involves processes from start to finish, including naming files, preparing and cleaning your data, performing analyses, documenting your work, and more. Below are selected resources help improve your workflows through better data management practices.

File Naming

• File names should be:
  • Human readable - you should understand the content of the file from its name alone
  • Machine readable - don't use spaces (_ or - are preferable), special characters or accents
  • Use default ordering - start with a leading zero (01, 02, etc) for numbered files, use the YYYYMMDD date format
• Have a plan for version control, either by including version numbers in the file names or using an automated system l

File Organization

• Always save raw data in a separate folder
• Make sure folder names are descriptive and consistent
• You can create a folder structure based on data type, processing stage, or any other system that makes sense to you and your collaborators

File Formats

• Be sure to save data in an open format for sharing and long-term preservation
• Check the Library of Congress Recommended Formats list to see what is recommended for your file type

Resources:

• MIT Libraries File Naming Handouts
• UK Data Service Guide to Data Organization
• Data One Primer on Data Management

Following a few basic recommendations when working with research data in spreadsheets can save you time when it comes to analyzing your data. Consider these practices adapted from Data Organization in Spreadsheets, Karl W. Broman & Kara H. Woo.

Basic Spreadsheet Practices

Getting Started

• Make backups
• No calculations in the raw data files
• Make it a rectangle and fill in each cell

Inputting Data

• Be consistent
• What to do with empty cells? Consider options as described here: DataONE - Data Entry and Manipulation
• Use data validation to avoid errors. For Excel, see: Excel Data Validation Guide | ExcelJet

Tidy Data

• Data Science for the Biomedical Sciences
  • Every column is a variable
  • Every row is an observation
  • Every cell is a single value
• Do not use font color or highlighting as data - consider an additional column with a "flag" value, as described above in  Exercise 3

Sharing

• Save the data in plain text files

Additional Resources:

Data Science for the Biomedical Sciences - Spreadsheets
Data Carpentry Spreadsheet Lesson
DataONE Data Entry and Manipulation (creating files, missing values, data validation)
Data Organization in Spreadsheets, Karl W. Broman & Kara H. Woo

Documentation:

Describing your Project

CESSDA has a useful guide for creating project-level documentation:

1. For what purpose was the data created
2. What does the dataset contain?
3. How was the data collected?
4. Who collected the data and when?
5. How was the data processed?
6. What possible manipulations were done to the data?
7. What were the quality assurance procedures?
8. How can the data be accessed?

Describing your Dataset(s)

Nice overview on Readme, Data Dictionaries, Codebooks with examples (Iowa)

Readme File

A readme is typically a plain-text file that provides information about a datafile to help facilitate use and re-use of the data. Typical elements to a readme include the following (adapted from Guide to Writing "readme" Style Metadata). Using one of the templates below can help ensure you create a useful readme file.

Readme Content: General Information

• Dataset title
• Creator name and contact information
• Date(s) of data collection
• Location of data collection
• Keywords to describe data topic

Readme Content: Data and Files

• Descriptive file names for each file, and for each, a description of what data is contained
• Data the file was created
• List of variables for each dataset, including full names and descriptions of each
• Definitions of any codes or symbols, including those for missing data

Readme Content: Methods

• Methods for data collection or generation
• Methods used for data processing

Additional Readme Resources

• Sample Readme file from deposited dataset "Data and R code to support: Estimating densities of zebra mussels (Dreissena polymorpha) in early invasions using distance sampling"
• LibraData (UVA's Research Data Repository) Readme template 
• Guide to Writing "readme" Style Metadata (including a template) (Cornell)

Data Dictionary

• How to Make A Data Dictionary (OSF)
• Describing Your Data: Data Dictionaries (Smithsonian)  

Codebook

• Nice, simple codebook example (Kent State) (part of a good SPSS tutorial on Creating a Codebook)

Metadata and Standards

Disciplinary Metadata (Digital Curation Centre) - links to information about metadata standards by discipline/field

• List of Metadata Standards
• List of Metadata Use Cases to look over for examples - e.g. ICPSR - Inter-university Consortium for Political and Social Research - a huge repository of datasets that have been professionally curated (described) with all metadata conforming to the DDI standard.

University Policies:

• UVA Data Protection Standards
• Additional UVA Data Policies

UVA-Contracted Cloud Storage:

• User responsibilites and data restrictions for UVA Box and OneDrive
• UVA Box
• UVA OneDrive

Additional UVA Storage Resources:

• Research Computing Data Storage
• ITS Storage, Hosting & Servers

Backing Up Your Data:

• Consider the rule of three:
  • Here (lab computer, personal computer)
  • Near (portable hard drive, flash drive)
  • Far (cloud storage, remote backup)
• CESSDA Guide to Backing Up Data

Archiving and Sharing:

• Institutional Data Repository
  • UVA LibraData
• List of Disciplinary Data Repositories
  • Re3Data
• Multidisciplinary Repositories
  • Figshare
  • Zenodo
  • Dryad
  • Open Science Framework

UVA Resources for Computing and Analysis:

• Research Computing
• ITS Software Gateway
• UVA Library Research Data Services + Sciences
• UVA Library Research Data Management Guide
• UVA Claude Moore Health Sciences Library Research Data Services

The Open Science Framework (OSF) is a free, open web-based platform to support your research and enable collaboration. Think of it as a site where you can manage your whole research project. See examples:

• OSF for a shared lab space, share lab standards and resources, and collate the work that is being done by individuals/groups within the lab
• template for researchers to use an OSF project as an Electronic Lab Notebook (Johns Hopkins University)

Project TIER seeks to enhance transparency and reproducibility of empirical research in the social sciences. Resources include: 

• The TIER Protocol (4.0) which provides a template for students to document statistical computation (e.g. folders for your input data, your analysis, your scripts, etc). A template exists in OSF for TIER.
• The DRESS Protocol is a set of standards for replication documentation that embodies the same principles that underlie the TIER Protocol but is tailored to suit the purposes of professional researchers

Need more information on managing your research data? We are here to help:

University Library Research Data Services +  Sciences - contact at dmconsult@virginia.edu 

Health Sciences Library Research & Data Services - contact at hsl-rdas@virginia.edu