Metaphors of Data

When scientists and practitioners try to explain matters about data, they very often refer to metaphors from the physical world. Most of the terms have been established long before the digital era, they come from commerce (“data storage”, “data retrieval”, “data mining” or “data harvesting”) and nature (“data explosion”, “data is the new oil”, “Datenberg” (in German)). Han-Teng likes to speak of “data massage”. He uses the term to describe the manual effort of getting raw data (!) into the right shape before it can be further processed.

The terminology of data is full of metaphors. And – as it lies in the nature of metaphors – they are never never precise, because the words are taken out of context, they stem from another sphere of meaning and should explain entities that are difficult to understand otherwise. For instance, the “new oil” comparison is inadequate because data is (usually) not a finite resource.

In need for a visual for data? A picture with ones and zeros is the worst to use.

In need for a visual for data? A picture with ones and zeros is the worst to use.

But explaining data without metaphors is also very hard. I usually speak about structured information or values in (MS Excel) tables. Video makers and news reporters also face that problem when they produce films on data issues. Usually they show the physical repositories of data: screens with visualizations, server farms and crowds on the streets. This rather new BBC documentary on big data is a perfect example for that. At the same time I find the animated data flows in the intro an appealing idea:

The human face of data

Very often, data metaphors are so common, we do not even realize they are metaphors. The trekkies here cerainly know an officer called Data. Until recently I did not associate Data with structured information. To me, Data was just a normal name. But Data is called Data because he is an android. As one of the leading officers on Picard’s Enterprise, he has impressive computational capabilities. He is also a know-it-all, more like an encyclopedia. But I guess that name would not have been so sexy. Take this example, where Data explains complex systems:

Data saves the enterprise on numerous occasions. His biggest problem is that he does not understand human behavior. He is unable to feel emotions. One thread that continues through the whole TV show is that Data tries to find a chip that would bring him human feelings. Until that is the case, Data is very slow on the uptake of getting jokes and habits:

Although Star Trek plays in the 24th century, Data is not able to differentiate between an idiom and a factual expression – something computers are already very close to understanding today.

In the following clip I found another note-worthy anthropomorphism of data. The video is an episode of a mini web show by Oracle to advertise their data processing solutions.

The video reminded me of The Internet Party, another short film, where services become humans to better understand their nature. Although very old, I still find that representation of persons can explain abstract concepts very well, especially when it comes to data.

Do you know other metaphors or visual representations for data? Please post them in the comments.

I gave this talk in a “Big Data, Small Data” YouTube Cinema that I organized during my fellowship at the Humboldt Institute for Internet and Society. On 11 September 2013 the other fellows and I showed videos and talked how data, big and small, interfered with our lives and our research.