Who was named,
and who was named for them
A reading of the 58-entry stakeholder long-list produced at the ENGAGED kick-off in Amsterdam — what the list shows about how the room saw the field, and what its structure leaves out.
An afternoon’s worth of naming, not yet a map
On 28 May 2026 the ENGAGED consortium and its three case-study cities met at the Hogeschool van Amsterdam for the project’s formal kick-off. The morning had run a World Café on the obstacles to equitable adaptation; the afternoon’s stakeholder mapping session, facilitated by Roberto Rocco, asked each city table — Amsterdam, Utrecht, Capelle aan den IJssel — to work through Parts 1 and 2 of the ENGAGED Stakeholder Mapping Canvas: name who is involved, then rate each name on influence, interest, and capacity.
The 58 entries analysed on this page are the consolidated Part 1 output. None carries the influence, interest, or capacity scores the session was designed to produce, and the per-city paper grids the agenda describes have been merged into a single sheet: only four of the fifty-eight names still carry the city they came from. What follows treats that gap as data in its own right, not as a footnote.
From the kick-off agenda
Four steps from paper grid to network
Identification
The raw output of the afternoon session: 58 names, each tagged by type (Community / Institutional / Professional), sector, and governance scale.
Consolidation
Three city grids merged into one sheet. City of origin was preserved for 4 entries; the other 54 became scale-only categories with no named organisation.
Attribute network
The session asked tables to map relationships and tensions between actors, but that relational data was not captured in the sheet. The network below is built instead from the three fields every entry shares — a similarity graph, not an observed one.
Reading against the canvas
The list is checked against Activity 4.1’s own equity-check criteria: vulnerable-group coverage, technical/non-technical balance, and coverage across governance scales.
Who got named, by sector, type, and scale
Institutional actors account for 48% of the list and government bodies for 38%; the private sector — the actors who assemble land, build housing, and price flood and health risk — accounts for 12%.
An attribute network, not an observed one
The spreadsheet records no relationships between named stakeholders — no reporting lines, no alliances, no declared tensions, even though the session’s own prompt asked tables to map them. The graph below connects two stakeholders only when they share all three recorded attributes (type, sector, and governance scale); colour marks the three clusters a Louvain community-detection algorithm found in that similarity structure. Drag a node, hover for detail, or recolour by a single attribute.
Community voice does not travel up the scale
Read by governance tier rather than by cluster, the list shows a clean structural pattern: community actors appear only at the local scale. No community-type entry was named at the regional or national tier.
What the equity check finds
Covered without prompting
Older residents appear twice (as a category and via a seniors’ association), and the table reached beyond the canvas’s own checklist to name gender groups, pregnant women, and low-literacy residents — categories the project brief does not enumerate but which sit inside the same intersectional-justice literature the project draws on. That breadth suggests the room did not simply default to familiar institutional actors, the risk the canvas itself warns against.
Conflated, not missing
Four of the canvas’s six vulnerable-group categories are present only by proxy, and the proxy is doing real conceptual work it should not be asked to do silently. “Migration background / labour migrants” merges long-settled residents with a migration background and labour migrants on temporary contracts — two populations with different relationships to housing precarity and heat or flood exposure. “People living alone” stands in for social isolation, which is a different thing. Low income and chronic illness have no entry at all, proxied or otherwise. Each of these should be split, or named as absent, before the list is scored on influence, interest, and capacity.
One entry the method cannot classify
“Animals” was named at the table and carries the identical type, sector, and scale profile as the human community entries, which places it in the same cluster as Ouderen and Kwetsbare bewoners in the network above. That is a property of the spreadsheet’s three-field ontology, not a judgement on the table’s reasoning: a more-than-human stakeholder cannot be scored on influence, interest, or capacity as the canvas defines them, and WP4 should decide deliberately whether and how to carry it forward rather than let the spreadsheet decide by omission.
What this page cannot tell you
- This is a similarity network, not a relational one. The session’s own prompt asked tables to map relationships and tensions between actors; the consolidated sheet records none. Any reading of “hierarchy” here describes formal Dutch multi-level governance (Rijk → Provincie/Waterschap → Gemeente), not an observed chain of command between the named entities.
- No Part 2 scores exist yet. Influence, interest, capacity, and the resulting Core/Consult/Inform/Monitor priority have not been assigned to any of the 58 entries, so nothing here should be read as a prioritisation.
- City provenance was largely lost in consolidation. Fifty-four of fifty-eight entries cannot currently be attributed to Amsterdam, Utrecht, or Capelle aan den IJssel specifically, even though the exercise itself ran as three separate city tables.
- Category labels stand in for named organisations. “Wijkmanagers” or “Politieke partijen” are roles, not contacts; turning this list into an engagement plan requires attaching a name to each.