Compare the six alternate compositions, 07–12.
Six generative studies of the air Detroit breathes. One pollutant, one year, one source: fine-particle pollution (PM2.5) measured at seven Detroit monitoring sites in 2025, from the U.S. EPA Air Quality System (AQS). The images translate concentration, time, location and gaps into tabular notation, traces and bars.
The subject has local significance: Detroit’s air-monitoring program and community training efforts make the air itself a matter of public attention. These visualizations do not identify the source of a pollution episode, assign blame to a facility, or infer anyone’s personal exposure.
The artworks remain numbered, with dates and compact numerical notation. This page provides the interpretation and units. The same light ground, condensed bold mono type and near-black ink run through all six. The two circular works are measured radial plots integrated with numerical registers. There are no decorative guide rings, stippled fills or simulated telemetry.
| Study | Mapping |
|---|---|
| 01 / Monitor register | Selected calendar month. Seven columns are the seven monitoring sites; rows are day of month. Each bold cell is the reported daily PM2.5 concentration in µg/m³. Site IDs, full-year sparklines and valid-day counts sit above the columns. A thin line tracks the selected day. 07 = site count; 2448 / 2555 = observed / possible site-days. Missing cells show an em dash. |
| 02 / Radial year | One large radial plot of Southwest Detroit’s 365 daily concentrations. Angle follows day of year; each spoke begins at a constant 300-pixel radius and extends by concentration / 57.4 × 150 logical pixels. The circle is formed entirely by measured sample marks; no extra outline, stipple or decorative rings are drawn. Numbers 01–12 mark month boundaries. Center figures give month/day, selected concentration and day-of-year / 365. The right register lists every daily value in the selected month with proportional bars. |
| 03 / Layered records | Seven daily station traces are drawn as a layered field, with exact point markers at reported observations. Horizontal position is date. Stations are offset by site-code order (21 logical pixels horizontally and48 vertically); this is a categorical separation, not geography or interpolated terrain. Concentration uses the same300-pixel/57.4 µg/m³ vertical scale for each trace. A thin spine joins reported observations on the selected date; it does not estimate values between stations. Missing dates remain gaps. Bottom registers give site ID, coordinates, selected concentration, valid-day count, and recorded min / mean / max. The mean is a descriptive mean of available daily values, not a regulatory annual design value. |
| 04 / Radial range | One radial plot showing the range across reporting monitors on each date. Angle follows the year. Each spoke’s endpoints encode that day’s minimum and maximum among reporting stations, on a shared radius: 300 + concentration / 57.4 × 135 logical pixels. The circular form is the record, not a modeled plume or exposure map. Center figures show selected min / max and reporting-site count / 07. Thin leaders link each reported selected-date value on the radial scale to its station register. Right-hand registers give selected site concentrations and coordinates; the left register shows nearby day-of-year and ranges. Missing station observations are excluded from the range, with the reporting count retained. |
| 05 / Ranked register | The 96 highest daily concentrations at Southwest Detroit in 2025, ranked descending, breaking ties by earlier date. Four columns of 24 rows. Each record gives rank, month/day, measured concentration and a proportionally scaled bar. The header 096 / 365 states the displayed subset. The small histogram above the list counts all365 primary-site observations in24 equal concentration bins spanning0–57.4 µg/m³; the last bin includes the upper endpoint. No AQI categories, source attribution or health threshold is inferred from rank. |
| 06 / Shared time | Three station records on one calendar: Southwest Detroit (0015), East 7 Mile (0019), Eliza Howell-NR (0093). All curves share a linear 0–57.4 µg/m³ scale (155 logical pixels). Missing dates stay as gaps. Large right figures are selected observations. The compact upper strip encodes daily reporting-site counts (0–7) across the year. Each lower numerical strip gives neighboring day-of-year / concentration; these strips wrap within the 2025 year. |
The typographic refinement draws on compact terminal resource-monitor layouts such as btop and aligned terminal tables such as Lip Gloss, alongside the visual references supplied for this project. The implementation remains original p5.js code; those projects are references, not imported dependencies. Body figures use regular mono weight, larger measured values use bold, and the same records drive every micrograph and leader.
EPA AirData / daily_88101_2025.zip, retrieved October 19, 2026. EPA file documentation. Filters: State Code 26; County Code 163; City Name Detroit; parameter 88101 (PM2.5 FRM/FEM); POC 1; Sample Duration “24-HR BLK AVG”; Pollutant Standard “PM25 24-hour 2012”; Event Type “None”.
The duration and standard select one reported daily block-average record instead of mixing it with the separate summaries of hourly measurements. The “2012” string is the agency’s aggregation identifier, not a claim about a current health threshold. Every selected source row in this snapshot has Event Type None. Values are used in their reported units: micrograms per cubic meter at local conditions. No AQI conversion, interpolation, synthetic observations, or averaging of duplicated summary rows is performed.
Selected coverage: 2,448 records out of 2,555 possible site-days, with 107 missing site-days. The Southwest site is complete (365/365); East 7 Mile has 323 days; Eliza Howell-NR 342; New Mt Herman Church 357; DP4th 363; Trinity St Marks 335; Military Park 363. These seven sites are the qualifying records in this file, not every monitor or sensor in Detroit. The source snapshot can be revised by its reporting agencies.
Missing values are stored as null and displayed as gaps or em dashes. Neighbor-sample windows wrap within the 2025 calendar. Cross-site ranges use only reporting sites and show their count. Tables, traces and bars are readings of the source records, not models of air movement. The work is a historical art study, not a live air-quality service or health advisory.
The six works cycle every 60 seconds through a four-second continuous crossfade. Left/right arrows or Page Up/Down navigate; swipe on touchscreens. Tap or move the mouse to reveal shared controls. Space holds playback, A toggles automatic cycling, F requests fullscreen, S exports a 3840 × 2160 PNG, C changes the restrained palette, Escape opens the grid. Reduced-motion devices start held and use cuts. Direct inspection: open tui.html?study=1 through 6 without show=1; click/drag selects a record, arrows step dates, and Backspace resumes the reading cursor.
Built with p5.js. Fonts are bundled DejaVu Sans Mono regular/bold with horizontal condensation for larger figures. All data are bundled locally; playback needs no live data request. The previous transit/weather/river records are no longer loaded by this collection.
Extracted records · Complete selected EPA rows · Manifest and source SHA-256 · Code, extraction script and source subset
Source records, units and timestamps are preserved for audit. No endorsement by EPA, the City of Detroit, EGLE, or Processing is implied.