The Museum's Method: What Curatorial Practice Teaches Us About Open Data
We often talk about open data and web archives in the language of engineering: harvesting, scraping, crawling, and storing. These are processes of accumulation, of capturing the digital torrent in ever-larger buckets. But what happens after the capture? This is where our language fails, and where an unexpected field—museum curation—offers a more meaningful vocabulary and a crucial set of lessons.
A curator does not simply gather every artifact from an era and pile them into a room. Such a hoard would be incomprehensible, a burden rather than a resource. Instead, curation is an act of thoughtful selection, context-building, and storytelling. A curator identifies what is significant, arranges objects to create narrative flow, and writes labels that transform an unfamiliar relic into a understood piece of history. The goal is not completeness for its own sake, but intelligibility.
Our approach to open data could learn from this. We frequently celebrate the sheer volume of a dataset—the number of terabytes, the billions of records—as its primary virtue. But a petabyte of raw, uncontextualized data is often as useful as a warehouse full of unsorted, unlabeled artifacts. The real value is unlocked not by the act of preservation alone, but by the subsequent acts of curation.
This means moving beyond the ‘save everything’ mentality. A museum acquires with intent, knowing its mission and the stories it aims to tell. Similarly, an archive of public records or a web collection should be built around a guiding thesis. Are we documenting the evolution of a public policy? Preserving a digital community’s culture? This intent dictates what is collected, how it is described, and how it is presented. It transforms a data dump into a curated collection.
The most powerful curatorial tool is the label. In a museum, a small placard tells you what an object is, why it matters, and its relationship to the items around it. In the world of data, this is the role of metadata. Rich, human-readable metadata is the placard for a dataset. It explains its provenance, its limitations, the quirks of its collection, and its relationship to other datasets. It is the difference between a folder of cryptic CSV files and a comprehensible public record.
Adopting a curatorial mindset shifts our purpose from mere hoarding to responsible stewardship. It asks us to be narrators, not just librarians. It forces us to consider not just if we can save something, but whether we should, and if so, how we can save it in a way that ensures it will be understood and useful decades from now. The goal is not to build a warehouse, but a museum—a place of learning, built with care and intention.
Notes & further reading
A few pages I came back to while writing this: