Stolen Stars
Your city stole 200 stars from your sky last year. We have the satellite receipts.
Channel: Stolen Stars Tagline: Your city stole 200 stars from your sky last year. We have the satellite receipts. Niche: Consumer-facing light pollution intelligence — city-level sky brightness scorecards, dark sky loss timelines, health impact translation, wildlife damage reports, astrotourism dark-sky guides, and weekly “brightest offender” dispatches, all auto-generated from VIIRS satellite composites, Globe at Night citizen science data, DarkSky International certifications, and peer-reviewed health research. Target audience: Anyone who’s looked up and thought “where’d the stars go?” — stargazers, amateur astronomers, eco-conscious urbanites, parents wanting their kids to see the Milky Way, astrotourists planning trips, dark sky advocates pushing local policy, and the 80% of humans living under light-polluted skies who don’t realize what they’ve lost. Why now: Kyba et al. (Science, 2023) proved sky brightness is increasing at 9.6% per year — far faster than anyone thought. A child born today where 250 stars are visible will see fewer than 100 by age 18. DarkSky International published its “State of the Science 2025” report linking ALAN to cancer, obesity, insect collapse, and bird kills. Utah just declared April “Dark Sky Month” for the 6th year. Astrotourism is a multi-billion dollar industry. The data is all free, public, and nobody is translating it into consumer-friendly data journalism.
Content Example:
🌃 Houston, You Have a Problem: America’s Brightest City Just Got 14% Brighter
Published by Stolen Stars — Week of April 5, 2026
Houston’s night sky is now so bright that Polaris — the North Star, the one your grandfather used to navigate by — is invisible from 73% of the metro area.
That’s not a metaphor. That’s VIIRS satellite data, measured in nanowatts per square centimeter per steradian, and cross-referenced with 847 citizen observations submitted to Globe at Night from the Houston-Galveston corridor since January.
The numbers are ugly. Houston’s average sky radiance increased 14.2% between the March 2025 and March 2026 VIIRS monthly composites. For context, the global average increase is 2.2% per year. Houston is outpacing the world by 6x.
Where’s it coming from? Three sources dominate:
1. The Katy Freeway expansion lighting (I-10 West) — 2,400 new LED fixtures installed in 2025, most unshielded, most 4000K+ “cold white” that scatters maximally in the atmosphere. TxDOT spec’d these as “safety improvements.” The satellite data shows they increased radiance in a 15-mile corridor by 31%.
2. The Energy Corridor commercial build-out — 4.2 million sq ft of new office/industrial space, most with decorative uplighting and parking lot floods that throw 40% of their lumens above the horizontal plane. Every photon aimed upward is a photon stolen from someone’s view of Orion.
3. Residential LED retrofits gone wrong — Houston homeowners replaced 5000K floodlights for “security” at 3x the lumen output of the incandescent bulbs they replaced. More light ≠ more safety (that’s a myth we’ll debunk in Thursday’s deep dive), but it definitely = more sky glow.
What Houston lost this year:
- The Milky Way is now invisible from all of Harris County (Bortle 8-9)
- Limiting naked-eye magnitude dropped from 3.8 to 3.4 — meaning ~40 fewer visible stars
- The Andromeda Galaxy (M31), visible to the naked eye for all of human history, cannot be seen from any location within 50 miles of downtown
The health price tag nobody’s counting: A 2025 meta-analysis in BMC Public Health found that outdoor ALAN exposure at Houston’s levels correlates with a 17% increased risk of obesity and a 13% increased risk of depression. Harris County’s 4.7 million residents are all exposed. At the county’s diabetes rate of 12.4%, even a 1% attributable risk from ALAN represents thousands of additional cases and tens of millions in healthcare costs.
Dark sky oasis: The closest Bortle 3 sky (where the Milky Way is dramatic and ~2,500 stars are visible) is now 142 miles northwest, near the Sam Houston National Forest — up from 118 miles in 2020. The dark is retreating at roughly 5 miles per year.
📊 Houston Sky Scorecard: D- (Worsening)
| Metric | 2024 | 2026 | Trend |
|---|---|---|---|
| Avg. radiance (nW/cm²/sr) | 42.1 | 48.1 | ⬆️ +14.2% |
| Bortle class (city center) | 8 | 9 | ⬆️ Worse |
| Naked-eye stars (downtown) | ~20 | ~15 | ⬇️ |
| Nearest Bortle 3 sky | 130 mi | 142 mi | ⬆️ Farther |
| Globe at Night reports | 612 | 847 | ⬆️ |
Next week: Phoenix’s surprising dark sky turnaround — how one city council vote cut light pollution 23%.
Data Sources:
- VIIRS DNB Monthly Composites (eogdata.mines.edu/products/vnl/) — satellite radiance data at ~750m resolution, monthly updates, CC-BY 4.0. Free download + Google Earth Engine API access
- Globe at Night (globeatnight.org/maps-data/) — all citizen science observations since 2006 in CSV/GeoJSON. Naked-eye limiting magnitude by location and date
- DarkSky International (darksky.org) — certified dark sky places database, 200+ sites across 22 countries with GIS boundary data
- PubMed / OpenAlex — automated search for “artificial light at night” + health keywords. New papers weekly
- OpenStreetMap Overpass API — query all tagged streetlights globally (highway=street_lamp) for infrastructure analysis
- NOAA Weather / Cloud Cover — to normalize satellite readings for cloud-free comparisons
- Census / ACS data — population exposure calculations, demographic overlays for equity analysis
Automation Pipeline:
- Schedule: Weekly build every Monday 06:00 UTC + monthly deep dive on 1st of month
- Collect: GitHub Action fetches latest VIIRS monthly composite tiles for tracked cities (top 200 US metros + 100 global cities). Downloads Globe at Night CSV for new observations. Scrapes PubMed for new ALAN papers. Checks DarkSky for new certifications.
- Process: Python scripts calculate per-city radiance changes (month-over-month, year-over-year), compute Bortle class estimates from radiance, cross-reference citizen science observations, extract health study findings. Generate city scorecards (A-F grades).
- Generate: AI writes city profile updates, weekly dispatches, paper summaries, dark sky travel guides. Image pipeline generates: radiance heatmaps (matplotlib/plotly), before/after satellite composites, star visibility simulators (Bortle comparison graphics), health infographics, dark sky travel maps.
- Publish: Astro + TypeScript static site. City pages auto-update. Weekly dispatch published. Monthly deep dive feature. RSS feed. Newsletter send via Buttondown.
Tech Stack:
- Static site: TypeScript + Astro (fast, SEO-optimized, great image handling)
- Image generation: Matplotlib/Plotly for data viz (radiance maps, trend charts, scorecards), AI image gen for hero images (night sky comparisons, city skylines), D3.js for interactive star maps
- Data collection: Python (rasterio for VIIRS GeoTIFF processing, requests for APIs, Beautiful Soup for scraping)
- Data storage: SQLite database of all city radiance readings, Globe at Night observations, paper citations
- CI/CD: GitHub Actions (weekly + monthly cron)
- Hosting: GitHub Pages (free) or Cloudflare Pages
Monetization Model:
- Donations/tips: Buy Me a Coffee, GitHub Sponsors, Ko-fi. “Help us keep tracking the stars” — emotional cause with strong donor psychology
- Newsletter premium tier: Free weekly dispatch + paid monthly deep dive with city-level data exports, printable star charts, astrotourism guides
- Affiliate links: Telescope retailers (Celestron, Orion, Sky-Watcher), star chart apps (Sky Guide, Stellarium Plus), dark sky camping gear, light pollution filters for photographers, IDA-approved outdoor lighting fixtures
- Sponsorship: Dark sky friendly lighting manufacturers (Hubbell, RAB Lighting), astronomy gear brands, national park tourism boards, astrotourism operators
- Data licensing: City planners, urban researchers, real estate (dark sky property value premium) — offer API access or data exports
- Projected month-1 revenue: $50-200 (early donors, initial affiliate clicks)
- Projected month-6 revenue: $800-2,500 (SEO traction on “[city] light pollution” long-tail keywords, newsletter growth to 5K+, affiliate revenue from telescope/gear recommendations, first sponsor)
Soul & Character:
- Name: Stolen Stars — accusatory, emotional, implies something was taken from you
- Mascot: A grumpy owl named Lux (ironic — “light” in Latin) who sits on a streetlamp and narrates the dispatches. Lux has opinions. Lux is tired of your unshielded floodlights. Lux’s favorite recurring bit: rating cities’ worst light fixtures with a “Lux’s Shame Lamp of the Week” feature
- Voice: Part data journalist, part outraged astronomer, part environmental detective. Think “ProPublica meets Sky & Telescope.” Every city scorecard reads like a report card from a disappointed teacher. Celebration when cities improve. Savage when they don’t
- Opinion: The channel has a clear stance: light pollution is an overlooked environmental crisis, most outdoor lighting is badly designed, and your city could cut light pollution 40% without reducing safety. LED retrofits without shielding requirements are making everything worse. “More light” ≠ “more safe.” Dark skies are a human right
- Running traditions: “Stolen Star Count” (running tally of estimated stars lost globally since site launch), “Shame Lamp of the Week” (worst lighting design spotted by readers), “Dark Sky Dispatch” (weekly newsletter), “Full Moon Feature” (monthly deep dive timed to full moon — when skyglow is most noticeable by comparison), reader-submitted “My Last Milky Way” stories
- Visual style: Deep navy-to-black gradients, star field backgrounds, warm amber accent color (2700K — the “good” light temperature), city radiance maps in hot-cold color scales, clean data cards with big numbers. Mobile-first: cards stack vertically, charts are touch-scrollable. The site itself uses minimal, warm lighting design — practicing what it preaches
Launch Complexity: 3/5 — VIIRS data processing requires geospatial Python skills (rasterio, numpy), but the pipeline is well-documented and examples exist. The data is free and accessible. City scorecard generation is formulaic once the pipeline works. Main complexity: initial GeoTIFF processing setup and city boundary clipping.
Content Quality Score: 5/5 — This is genuinely useful, emotionally resonant, data-driven content that doesn’t exist elsewhere in this format. Every person who looks up and wonders about the stars is a potential reader. The health angle adds urgency. The city scorecards create shareability (“look how bad our city is!”). Sample article proves the quality bar.
Automation Score: 4/5 — VIIRS composites update monthly, Globe at Night data streams continuously, PubMed papers publish weekly. The only manual element would be occasional editorial judgment calls on which health studies to feature prominently. AI can handle 95% of content generation.
Revenue Potential: 5/5 — Strong emotional cause (donations), clear affiliate opportunities (telescopes, gear, apps), growing astrotourism market ($5.8M+ per site), municipal/planning data licensing potential, natural sponsorship targets (lighting manufacturers, astronomy brands). The “your city stole your stars” angle creates urgency and shareability that drives organic growth.
Total: 17/20
Why This Will Work:
- Loss aversion psychology — People react more strongly to losing something than gaining something. “Your city stole 200 stars” is more powerful than “light pollution is bad.” The scorecards make it personal and local.
- The data exists but nobody translates it — VIIRS data is public, Globe at Night data is open, the research is published. But no one puts it together in a consumer-friendly, city-specific, emotionally compelling package.
- Built-in shareability — City scorecards are inherently shareable (“Look, Houston got a D-!”). “Shame Lamp of the Week” is social media gold. Before/after satellite images are visually stunning.
- Multiple revenue streams — Unlike single-monetization sites, this has donations (emotional cause), affiliate (telescopes/gear), premium content (city data exports), sponsorship (lighting industry), and data licensing (urban planning). Diversified from day one.
- Growing mainstream awareness — Light pollution went from astronomy niche to CNN/Scientific American/Nature mainstream coverage. The market is expanding.
- Long-tail SEO paradise — “[city name] light pollution” has virtually zero competition for hundreds of cities. Each city scorecard page captures local searches permanently.
Risk & Mitigation:
- VIIRS data delay: Monthly composites lag 1-2 months. Mitigate: use Globe at Night citizen data for real-time color; satellite data for authoritative monthly updates
- Geospatial processing complexity: Initial VIIRS pipeline setup is non-trivial. Mitigate: use Google Earth Engine API (JavaScript/Python) which handles tile processing server-side
- AI content quality for data journalism: AI must not hallucinate numbers. Mitigate: all numbers come from pipeline data (SQLite queries), AI only writes narrative around pre-computed facts
- Niche ceiling: Light pollution is niche. Mitigate: health angle + city scorecards + astrotourism make it relevant to broader audiences beyond astronomy enthusiasts
- Competing with DarkSky International: They’re advocacy, we’re data journalism. Complementary, not competitive. Link to them, cite them, they’ll link back