Headlines and coverage summaries are AI-generated and may contain errors.

Reports

Seeing change is hard until you zoom in.

There’s the famous story about the frogs. Drop them in hot water, they’ll jump right out. Turn up the heat slowly and they’ll stay in there until they boil.

It’s false, but it captures a real fear. Change, for both better and worse, can be hard to see in the moment. What, we have corruption now? We’ve always had it. Racism exists? I’m shocked.

So how do we test our impression that things have changed with data? The idea of this site is to take one random news day from one presidency and compare it with the parallel day (weekday locked) of another one. With just one day laid next to another, differences become apparent. I can tell you what I see here, in these comparisons, but I think I’d rather just leave it to you, the user, to look at it yourself and come to your own conclusions.

Look at a random
Biden–Trump day

Look at a random day, then walk through a couple more days to get a sense of any shifts using the forward arrows. Read all the items of the day, not just the top ones.

Summaries, categorizations, and deduplication are all done using AI over large datasets of events; expect a level of error.

← Introduction

Search on an issue

Browse the AI-assigned concern tags. Choose one to see its coverage across the archive.

Loading issues…

Summaries, categorizations, and deduplication are all done using AI over large datasets of events; expect a level of error.

Statistical Reports & Shift-Share Decomposition

Compare event volume, daily rates, rate ratios, and structural composition shifts across all five presidential terms (632 matched days in office per term; 31,476 total events). Click any category row to inspect its underlying headlines and news pegs.

Loading 31,476-event analytics cube…

Summaries, categorizations, and deduplication are all done using AI over large datasets of events; expect a level of error.

Joe Biden

Loading…

Donald Trump (T2)

Loading…

Keep browsing

Comparison settings

Dates

Event

AI Summary of Coverage

Possible Tags (auto-applied)

Potentially related coverage

Suggestions sharing specific names or subjects.

Stories with similar tags (auto-applied)

Matches require 80% of the first 6 topic, institution and law tags, then 30% of additional tags, plus at least one concern tag and at least 3 tags overall. Extra shared concerns count triple.