It’s probably mostly a coincidence that the asinine media content that passes for political campaigning arrived at my eyeballs around the same time Steve Ballmer did. But his sincere effort to improve our nation’s discourse invites—nearly compels!—my commentary on how things could be. You see, I know how things could be. We could have a much richer data society.

Being in a politically competitive state still feels new. Living in California and Washington, DC, for the first 50 or so years of my life kept me off the receiving end of American political campaigns. Nobody spends precious campaign dollars in these non-competitive jurisdictions.

But New Hampshire is different. There, during recent months, your cell phone buzzed every 10 minutes to let you know that Chris Pappas dated an Amazon lobbyist. Every YouTube video came prepended with the fact that John Sununu made “up to” $35 million making healthcare more expensive.

The chain of inferences from such allegations to useful policy or political knowledge is about the distance from New Hampshire to the Golden State. Walking.

It was in the heart of the campaign season that USAFacts introduced a feature called “Debate,” opening it with a discussion between Scott Galloway and Ben Shapiro on the American Dream. Introduced by Steve Ballmer and moderated by PBS NewsHour’s Geoff Bennett, it is a fine thing for what it is. No decent person could object to it. Here goes.

Our nation’s political discourse is not much improved by having two smart people discuss gauzy concepts like the American Dream. Relative to what’s possible, such debates are themselves rather gauzy.

“Relative to what’s possible” does the lifting there. Data should be helping to constitute a new “constitution of knowledge,” as Jonathan Rauch called it in an excellent National Affairs article and later a book of the same name.

We need a new constitution of knowledge because our institutions have not kept up with the public’s expectations and demands. Many institutions (such as news gathering and higher education) have proven partisan and rigid. Push-button access to useful information in other walks of life has made government entities and processes conspicuous in their opacity. The people don’t trust the system.

The USAFacts debate was fine. Really, it was fine! But aggregate data about homeownership or social mobility (elements of the Dream) are not likely to absorb into the body politic, even pored over for us by smart, interesting people. It reminded me of a subtle but important shift in the government transparency effort that opened with the first Obama administration.

Government data could draw connections between votes in Congress and money spent on roadbuilding and bomb dropping. Government data could tie regulatory enforcements and airport pat-downs to the elected officials who set them in motion. That would make for more meaningful campaigns and more informed voting—a healthier democracy.

In 2013 testimony to the US House Committee on Oversight and Government Reform, I described what had happened instead: a shift from “open government” to “open government data.”

The Obama administration called on agencies to release “high-value” datasets in 2009. I analyzed them, testifying later

In a grading of the data sets, I found that most failed to expose the deliberations, management, and results of the agencies. Instead, they provided data about the things they did or oversaw. The Agriculture Department produced data feeds about the race, ethnicity, and gender of farm operators; feed grains, “foreign coarse grains,” hay, and related commodities; and the nutrients in over 7,500 food items.

We thought we were going to get insider access to our government; we got the data the government holds about commodities, and about us. The USAFacts debate reminded me of that. Galloway, Shapiro, Bennett, and Ballmer were interpreting aggregate data about us for us.

My 2013 testimony cited and summarized years of work in this area, so it is good for a deeper dive. It has the recipe that I think the nation still hungers for.

I recommend that you focus your transparency efforts not on Web sites or other projects that interpret government data for the public. Rather, your task should be to make data about the government’s deliberations, management, and results available in the structures and formats that facilitate experimentation. There are dozens—maybe hundreds—of ways the public might examine the federal government’s manifold activities.

This is at an angle to the USAFacts effort. But I would like that organization to conceive of the problem differently, or consider a different problem. There is deep mistrust of our institutions. It won’t be solved by data per se, but through data that actually opens one of our most important and opaque institutions up to the multifaceted and highly granular analysis you see, for example, in baseball.