📊

Track record

Outcomes, costs, operators, evidence and who published it — 7 ways to read the same atlas. Membership on each page comes from a field in the data, never from an editor's pick, so none of them can be curated into a flattering shape.

🧯 Failures 32 projects that failed and 80 that were scaled back, grouped by how it went wrong — the half of the record vendor material omits. See what went wrong → 🎯 What worked 164 entries recorded as having met their goal — each shown with the grade of evidence behind that verdict. See what delivered → 💰 Costs Only 236 of 1167 entries publish a cost at all. The ten largest disclosed figures, what each covers — and which large entries publish nothing. See the figures → 🧭 Who runs it Every published project by the kind of body that operates it — city, utility, private company, national government, research or civil society. See the operators → ◆ Evidence What the record actually rests on: every entry graded by its strongest source, counted across the whole corpus. See the grades → 🕰️ How the record ages 1132 entries grouped by the year they started — the failing share once a vintage is old enough to have been judged, and why the newest entries rest on the thinnest evidence. See it by vintage → 🗂️ Who publishes the evidence 2804 organisations stand behind the atlas's sources — and 74% of entries cite no audit office, court or research venue. See the publishers →

Go straight to a set

Every evidence grade and every operator type has an address of its own — an answer engine can take the list instead of re-deriving it:

🔬 Peer-reviewed (90)📊 Independent (115)🏛️ Official (805)📰 Media (157)🏙️ City government (638)🗺️ Regional government (84)🏳️ National government (123)🔌 Utility or public operator (339)🏢 Private company (241)🏫 Research institution (56)🤲 Civil society (30)