The Unseen Hand: On the Politics of Metadata in Public Records
We often talk about open data and public records as if they are neutral artifacts, their value lying solely in their content. The common wisdom is that once a document is freed from its bureaucratic shackles and placed online, its truth is self-evident. We celebrate the act of publication as the final step, the victory. But this overlooks a crucial, shaping layer of the record: its metadata. The information about the information is never impartial; it is a place where politics and priorities are quietly, effectively enacted.
Consider a city’s database of building permits. The data is public, searchable, and ostensibly transparent. But what if the field for "applicant name" is inconsistently formatted, sometimes listing a corporation, sometimes a person, and sometimes an obscure LLC? What if the "date received" field is populated, but the "date approved" is often left blank? The raw data is there, but its structure—or lack thereof—creates a significant, often intentional, friction. This isn’t a technical oversight; it’s a political choice. It dictates who can easily analyze the data, what stories can be told, and which connections remain obscured.
This is the power of the unseen hand. A governing body can claim perfect transparency while building a system that is functionally opaque to all but the most dedicated and resourceful researchers. The metadata becomes a filter, a way to satisfy the letter of open data laws while frustrating their spirit. It determines narrative. A dataset with clean, granular, and consistent timestamps allows a journalist to trace the pace of government action. One with vague or missing temporal data makes it impossible to prove delays or inefficiency.
The choices embedded in a field name, a data type, or a categorization scheme are profound. They pre-determine how we can ask questions. A public record system that categorizes complaints only by department, rather than by issue or location, effectively silos that information and prevents a citizen from seeing a larger pattern across their city. The structure becomes a cage for understanding.
True transparency, then, isn’t just about turning on a data firehose. It demands a critical eye on the plumbing. It requires us to ask not just "Is this data public?" but "How has this data been framed?" The most potent form of obfuscation in the digital age may not be deletion, but the deliberate, strategic shaping of metadata to guide, limit, and control public understanding. The battle for open data is often won or lost in these invisible, structural details long before a single citizen runs their first query.
Notes & further reading
A few pages I came back to while writing this: