Method, and what it does not prove.

DWSD selects alleys for the $184M Alley Sewer Repair Program using closed-circuit camera inspections that confirm failed private connections, recorded alley cave-ins and sinkholes, and federal rules requiring block groups above 50% low-to-moderate income. Work is rolled out in phases, and residents cannot apply.

In a 2025 public comment response the City stated it already holds over 30,000 data points showing where failed lateral connections are.

I queried all 782 feature services on the City of Detroit’s ArcGIS organization. For sewers, exactly two layers are public: catch basins, and gravity mains with recent cleaning work orders. None of the 30,000+ failed-connection points are published. Neither is the list of selected alleys.

So I used the criteria that do have open data behind them:

• Alley cave-ins and sinkholes, Improve Detroit 311, 2026-09-18 snapshot
• Water-in-basement reports, same source
• Low/moderate income share, HUD by block group, live
• Council district, City of Detroit 2026 boundaries
• Alleys already contracted, DWSD Capital Improvement Projects, phases Construction and Procurement

I took the 138 alley projects in the current program (65 in construction, 73 in procurement) and the 47 alley projects the City completed before it, and measured 311 density within 500 m of each midpoint.

A permutation test on the difference of medians gives p < 0.0001 for the basement-flooding gap and p = 0.0115 for the cave-in gap. Across all 14,115 basement-flooding reports, 1.0% fall within 200 m of a contracted alley and 7.2% within 800 m.

The plain reading: the selection tracks collapsed pavement, which is a City asset liability it can see, rather than flooded basements, which residents report and which may never become a DWSD record.

BSEED trades permits record private sewer work by address: 2,722 permits since 2019, of which 1,288 are backwater valve installations, households paying to protect themselves. I measured the same 500 m flooding density around those addresses.

Two independent comparison groups, private spending and the City’s own earlier alley work, both sit at 38 to 45. The new program sits at 18. That makes the gap harder to explain as noise in 311 reporting: the same data, measured the same way, puts everything else roughly twice as high.

A permit also reflects who can afford to pull one, so this is not a clean measure of need. It is a third vantage point that points the same direction.

The public sewer dataset is DWSD’s Sewer Cleaning Dashboard, mains that carry a recent cleaning work order. It is a record of maintenance activity being read as an asset map, and it is thin where it matters.

I sampled 1,500 of the basement-flooding reports at random and asked whether any city sewer main appears within 120 m in the public data. 432 did, 29%. For the other 71%, a resident looking up their own block finds nothing at all.

In May 2026 a Detroit couple told a TV station about fifty years of basement flooding. DWSD’s deputy director said on camera: “This sewer is not on our records, it’s not on our maps.” The city cleared the line within days of the broadcast. A gap in the published record is not proof that DWSD lacks the data internally, but it is why a household cannot check anything for itself, and it is why this report says “no record” rather than inventing a pipe.

Four of seven districts have nothing in the first round, including District 4. There are three more years of the program, so this can change, and that is the reason to publish it now rather than afterwards.

• This is the first round. 138 alleys against roughly 9,000 connections planned over 4 years. Field work starts October 2026.
• The camera evidence is invisible to me. DWSD may have sound engineering reasons I cannot see. My report says so on every address.
• 311 records are requests. Not confirmed incidents, not unique households. They reflect who calls the City, which is itself unequal.
• The comparison group is small, 47 earlier alley projects.
• This is correlation, not an audit. It is a reading of public data, offered so that residents and the City can check it.

The calibration is one script, scripts/calibrate-asrp.mjs , run against the bundled snapshots, and it writes the numbers this page displays. The repository is open source.

Sources: DWSD Alley Sewer Repair Program · $184M announcement · Capital Improvement Projects · Improve Detroit 311 · HUD low/mod income · full list