Did council side with the room?
For each issue with eight or more speakers on a side, the council's actual action, set against what most speakers asked for. Sourced from minutes and news; each row links to the record.
Which districts come to the mic
Speakers placed in their home council district when they said where they live or described their own street or neighborhood. Each person counts once, however many times they spoke.
What speakers wanted done
Each speaker's main request, sorted into one type.
Show as a table
The biggest fights
Recurring issues, counted across every way speakers and clerks described them. Select one for what it was and examples of what people asked.
When people showed up
Speakers per month. Hover a bar for the issue that drove it.
Search what people asked for
One row per speaker, newest first, paraphrased from the transcript. Residents are not named.
Method and coverage
- Source. Each city's own minutes and speaker lists, plus the transcript or captions of the meeting video (Swagit, Granicus, the city clerk's closed-caption files, or YouTube). A few meetings with no captions were transcribed from the audio.
- What counts. Members of the public at full council meetings, both open public comment and speakers on specific agenda items. Committees, boards, commissions and briefings with no public comment are left out, as are registered speakers who did not show up.
- Coding. Each speaker's remarks were read and coded for a main topic, the kind of ask, a one-line paraphrase of the ask, and who it was aimed at. Recurring issues were then grouped across all the ways they were described, so each issue sits under one topic.
- The chart. Each named issue was framed as one plain yes-or-no question, such as "Should El Paso keep its incentive agreement with Meta's data center?", and every speaker on it was re-read and marked for, against or neither. Issues with at least five speakers on a side are plotted. Before placing a dot, five speakers at an even split are added, so an issue with six speakers can't sit at the edge while one with three hundred barely moves. Hover a dot for the exact counts.
- Heat. Each speaker's tone was coded from the transcript as calm, urgent, angry or appreciative. A dot's color is the share of its speakers who were urgent or angry. Tone is read from text, not voice, so sarcasm and inflection are lost.
- Mic vs. vote. For every issue with eight or more speakers on a side, the council's action was looked up in the minutes and news and mapped onto the issue's question. "Sided with the room" means the action matched what most speakers asked for. Issues decided by voters, the state or staff, or never voted on, are listed separately rather than scored.
- Districts. Boundaries come from each city's GIS service; population and household income from the Census Bureau's 2020–2024 American Community Survey, apportioned from tracts to districts (by area, or by tract center where that matched the city's equal-population districts more closely). Figures are approximate. A speaker is placed only when they said where they live or described their own street or neighborhood; organizations speaking citywide are not placed. People are counted once, using the minutes' speaker names where available. San Antonio and El Paso had too few placed speakers to compare districts.
- Scripted campaigns. A check for near-identical remarks at the same meeting found almost none (4 of about 4,700 speakers). Organized campaigns share talking points, not scripts.
- Speaker types. Residents and advocates includes nonprofit, business-group and union representatives. Applicants speaking on their own projects, officials of other governments, council members and city staff are excluded.
- Accuracy check. An independent re-read of 120 randomly chosen speakers found the topic wrong once, the kind of ask wrong in 4%, and the paraphrased ask only partly right in 13%. About 7% of transcripts were attached to the wrong speaker or ran several speakers together, mostly in San Antonio and Fort Worth.
- What it is not. Not a poll. It records who came to the microphone, which skews toward organized campaigns and repeat speakers.