DataFlash: Data Horror Stories

Posted by Ann Madhavan on February 15th, 2018 Posted in: Data Science
Tags: Data_Science

---

In the spirit of Love Data Week’s 2018 theme, Data Stories, it’s important to consider cautionary tales as well as good outcomes. We should, after all, learn from our mistakes. Perhaps the best known collection of data horror stories is Dorothea Salo’s Research Data Management Horror Stories pinboard. Dorothea, a University of Wisconsin academic librarian and library-school instructor, has been pinning data tales of woe since 2010.

We probably all have our own personal examples of data hell, but here are a few of my favorite themes…

• Submitting a grant proposal and neglecting to include a well thought out data management plan. Proposal rejected. Research flat lines.
• Gathering your identifiable biomedical data without adequately consulting with your Institutional Review Board (IRB). Collateral damage.
• Neglecting to develop and implement a detailed naming convention for your data files. Data hot mess.
• Neglecting to maintain thorough metadata for your datasets, models, and algorithms. Data bedevilment.
• Failing to sync and back up your data in three separate locations. Deadly data loss.
• Dismissing the need to follow guidelines to insure the security of your data. Data access nightmare.
• Saving your data to a proprietary file format that is on the verge of insolvency. Walking dead data.
• Disregarding the need to place your data in an appropriate repository that provides long term access and maintenance. Evil dead data.
• Facing the shame of having your publication retracted due to data irregularities. The horror. The horror.

For the love of all things organized, don’t let your researchers be condemned to these nine circles of data hell!