Network Matrices — Machine-Readable Graph Data

This page is the machine-readable layer of the network map. The Mermaid diagrams show the network visually. These matrices show it computationally — every relationship as a cell in a table, every money flow as a row in a ledger.

Download Formats

Board Membership Matrix ✅ Verified

Who governs what. Every cell is a public record. Blank = no documented role.

Pattern: Two convicted felons control the revenue entities (schools + CMO + foundation). Three judges + two government officials govern the boards. The felon and the judge sit on the same org chart.

Court Assignment Matrix ✅ Verified

Which judges sit where. The enterprise generates cases in Wayne County courts. These judges have documented connections to the enterprise.

Coverage: 4 judges in 3rd Circuit, 3 judges in 36th District, 1 in Probate, 1 in Oakland. Any case involving Banks’ network in Wayne County risks landing before a connected judge.

Money Flow Matrix ✅ Verified: All amounts from CFRS, TransparencyUSA, or audited financials

School Revenue Extraction

Campaign → Enterprise Payments

Enterprise Self-Dealing Loop

State per-pupil funding
    → Schools (PCA + MacDowell)
        → Purpose Group LLC (72.67%)
            → Banks (sole member, salary + fees)
                → Banks Strategy LLC (consulting entity)
                    ← Judge campaign payments ($383.82+)
                        ← Judges who hear enterprise cases
                            ← Cases generated by enterprise schools

Family Matrix ✅ Verified

Institutional Conflict Matrix ✅ Verified

Where personal roles conflict with institutional duties.

For Analysts and AI Agents

Python (NetworkX)

import json, networkx as nx

with open('graph.json') as f:
    data = json.load(f)

G = nx.DiGraph()
for n in data['nodes']:
    G.add_node(n['id'], **n)
for e in data['edges']:
    G.add_edge(e['source'], e['target'], **e)

# Find all paths from state funding to Banks
for path in nx.all_simple_paths(G, 'macdowell', 'banks'):
    print(' → '.join(path))

# Centrality — who is the most connected?
cent = nx.degree_centrality(G)
for node, score in sorted(cent.items(), key=lambda x: -x[1])[:5]:
    print(f"{G.nodes[node].get('label', node)}: {score:.3f}")

R (igraph)

library(igraph)
edges <- read.csv("https://detroit.primals.eco/graph.csv")
g <- graph_from_data_frame(edges[, c("source_id", "target_id")], directed = TRUE)
plot(g, vertex.label = V(g)$name, vertex.size = degree(g) * 3)

Gephi

1. Download graph.csv
2. Import as Edge Table (source_id → target_id)
3. Use edge_type for edge coloring, weight for thickness

All data derived from public records. Graph data is CC-BY-SA-4.0. Clone the repository for the full evidence package.