Dec. 3 Presentation Explores Data Disasters, Lessons Learned

The oldest mistake on a “spreadsheet” dates back 5,000 years to a Sumerian tablet in Mesopotamia. It’s fitting that one of the first civilizations in the world carried the earliest example of human error. The tendency to err has plagued us from the very beginning.

“We have been making mistakes since we learned to write numbers,” said WSU Libraries’ new science librarian Emily Cukier. She will give a presentation on the subject, titled “Data Disasters: Lessons from Notorious Mistakes in Spreadsheet History,” at 1:30 p.m. Friday, Dec. 3, in Spark 212. “We will never be mistake free, and that’s a liberating thing. It frees people up to not strive for perfection. That said, we have to try to catch errors before they happen.”

Data is everywhere, and data literacy is crucial to interpreting information received in school and over the course of our lives. Interested in data management since library school at Dominican University in River Forest, Ill., Cukier said it’s important to empower students to become data scholars and to contribute their own diverse perspectives to the scholarly conversation.

“Scholarly conversation is becoming very democratic, but there are still gaps in who can contribute to and manage data,” she said.

What bad data teaches us

A more recent example of a stupendous data disaster occurred in 1999 when NASA’s Mars Climate Orbiter went silent during its mission to Mars. The problem? A navigation error caused by a failure to translate English units to metric. The orbiter likely burned up in Mars’ atmosphere after it missed its target altitude due to the faulty information, costing NASA roughly $325 million, not to mention the years of science that went into the mission.

“The moral of the story is to think about who else wants to use the data and what they need to know to use it properly,” Cukier said. “We need to build systems that prevent mistakes rather than relying on individuals to be perfect.”

Another common mistake is improper data storage, she said. Lost or corrupted data cost time and money.Best practice calls for saving three copies of data in two different forms and one offsite—or the 3-2-1 rule.

Finally, she suggests that people get more comfortable about making mistakes. No matter how bad the screwup, it’s rare that a homework or research mistake will have severe consequences.

“Mistakes are made to be learned from,” she said. “They’re not sources of shame.”

With undergraduate and graduate degrees in chemistry and biochemistry, respectively, Cukier has much to lend to her role as a data librarian. She also worked in the private sector, first as director of chemical research for branding giant Brand Institute Inc. and then as senior writer for biopharmaceutical research company BioCentury Inc. She was drawn to WSU partly because her father teaches at another land-grant institution, Michigan State University.

Cukier said she hopes to help students develop their skills for managing data. This may include workshops and library instruction on topics like keeping spreadsheets clean, recording and preserving data for reuse, and interpreting data found within and outside of academia.

“Being comfortable with collecting and using data benefits students across all disciplines and will serve them well beyond academia in our increasingly data-saturated world,” she said. “Developing a skill for finding good data crosses many platforms and benefits multiple disciplines, not just academia.”

—Story by Nella Letizia