The Ghost in the Spreadsheet: On the Unarchived Voices of Public Comments

I spent a week last summer inside a PDF. Not a normal PDF, but one of those monolithic, thousand-page documents that constitutes the ‘public record’ for a local zoning proposal. My mission was simple: find the public comments. I wasn’t looking for the final tally of ‘for’ and ‘against,’ but for the voices themselves. I wanted to read the stories, the fears, the hopes that people had poured into the digital suggestion box.

What I found was a ghost town of data. The record contained a perfectly formatted spreadsheet. Column A: ‘Commenter ID.’ Column B: ‘Received.’ Column C: ‘For/Against/Neutral.’ Column D, the kicker: ‘Comment (Summary).’ Each unique, messy, human paragraph had been processed, summarized, and sanitized into a single, bureaucratic sentence. ‘Expressed concern about traffic congestion and loss of green space’ was the entire record of what was likely a passionate, personal plea from a lifelong resident. The raw data, the actual text of the comments, was not included. It was deemed, I suppose, too unwieldy, too unstructured, too human to archive.

This is the quiet erasure that happens in plain sight. We talk about link rot and format obsolescence as the great enemies of digital preservation, and they are. But this was different. This was a choice. The data was ‘open’ in the most technical, checkbox sense. You could download the PDF. You could count the responses. But the meaning, the substance, the collective voice of a community had been algorithmically distilled into a set of metrics. The record was preserved, but the testimony was lost.

It reminded me that the most crucial layer of preservation isn’t always the bitstream or the file format; it’s the context and the content in its original, unvarnished form. A summary is an interpretation. It filters out the quiver in a sentence, the local idiom, the unexpected tangent that reveals a deeper truth. By archiving only the summary, we don’t preserve public sentiment; we preserve a bureaucrat’s or an algorithm’s CliffsNotes version of it.

This moment cemented for me that true open data and readable records must fight for the raw material of civic life, not just the polished reports. The unreadable record isn’t always one that’s encrypted or corrupted. Sometimes, it’s one that has been meticulously, politely, and efficiently stripped of its soul, leaving behind a hollow and compliant ghost. Our work is to argue not just for access, but for authenticity—to ensure the ledger records the words as they were spoken, not just the boxes that were checked.

Notes & further reading

A few pages I came back to while writing this: