01
Required inputs
Policy context
Urban passenger mobility
Private car · Public transport · Active mobility
ReadyChange any input to recalculate
02
Target 2031
Initial city modal split
Modal Split Target Changes till Target Year
Default scenario target changes relative to the initial city modal split. These values are adjustable.
Baseline modal split Monday–Sunday expected shares
03
Editable assumptions
KPI impact & SOMI settings
Daily modal draw→Daily mode trips vs selected baseline→Dynamic KPI values→Normalised KPI / pillar scores→Persona SOMI→Solution SOMI
Impact coefficients Editable KPI response per +1% mode trips versus baseline
Persona–pillar & solution–persona weights Editable D1.4 defaults
The selected solution’s KPI values, normalization references, operational-impact inputs and both weighting layers are loaded and adjustable. Dynamic raw KPI values may move beyond a bound; their normalized scores are clipped to 0–1 while raw values remain available in exports.
A
Daily SOMI by persona
Daily simulated persona scoresEach marker is one simulated day
B
Nine persona-specific graphs
Weighted pillar lines versus persona SOMI for the selected solution
C
Static comparison, not the daily stochastic result above
How persona modal-split behaviour affects SOMI
Tool 3 · Scheduled semantic feedback → SOMI
Tool 2 supplies the daily simulation. A scheduled GO-DATA and semantics response, followed by Gemini KPI mapping and review, adjusts normalized KPIs during its reference period. The dashboard compares the resulting SOMI with Tool 2 at every level.
Start here · selected city example
Choose Florence, Pilsen, or Antwerp above, then run the example. Antwerp uses the attached 45-comment GO-DATA response with a provisional ChatGPT allocation. Florence and Pilsen use clearly synthetic comments and ChatGPT-authored allocation examples for every solution. Each city run calculates Tool 2 behaviour, applies the example semantic feedback, and aggregates all its solutions to City SOMI.
No credentials or installation needed for these examples.Scheduled API pipeline
Scheduler→GO-DATA sentiment GET→Comment JSON→Gemini KPI mapping JSON→Reviewed KPI adjustment
GO-DATA v1 docs: the server obtains a token from /auth/token and reads GET /v1/sentiment and GET /v1/sentiment/{post_id}. For fresh GO-DATA and Gemini extraction, use the integrated Flask backend. The URL below is local to that computer; opening the HTML alone does not start it. Set the actual client ID, secret and Gemini key in the backend, never in this page.
Triggering a run reads the selected city's tracked reports, retrieves one post, and requests Gemini KPI mappings. A post ID selects a specific tracked post; otherwise the latest analyzed city post is used. The backend keeps a daily or weekly schedule after the first run. Manual JSON imports remain available.
Check the integrated Flask backend connection. Manual JSON import and the supplied example work without the backend.
SOMI before and after semantic feedback
Choose a pillar, persona, solution or city and inspect the reference period. City SOMI appears when all Tool 2 solutions for the selected city have been finalized.
Semantic JSON and pillar–KPI allocation
Run the supplied example or import GO-DATA and KPI allocation responses.
View extracted GO-DATA sentiment JSON
View KPI allocation JSON and justifications
Pillar–KPI allocation used in the scenario
The table shows selected comments (or eligible mappings before selection). Excluded comments and their reasons remain in the raw allocation JSON and the full comment audit. Review sentiment and relevance before research use.
Import the responses to display allocations and reasons.
City modal split → persona behaviour → semantic SOMI
The daily city modal draw is shared by all personas in the reference stage. Tool 2 uses each persona's own day-type mode factors and stochastic draw. The semantic stage changes mapped normalized KPIs and retains the persona's Tool 2 mode choice. Tables show selected-period means. Different solutions may have different mode catalogues, including Florence freight.
Selected solution: city split and each persona's mode choice
All solutions and City SOMI
Selected solution: pillar, persona, and solution SOMI
Inspect every persona × pillar score
Which pillar KPIs changed?
Reference-period daily mean. KPI scores and pillar scores are normalized from 0 to 1. The pillar SOMI contribution applies each persona’s pillar weight and solution influence weight; the eight contributions sum to solution SOMI. The before values include Tool 2 behaviour, and the after values include the selected semantic feedback.
Run the example or apply imported mappings to see KPI and pillar changes.
Impacted KPI scores
Pillar scores and SOMI contributions
Calculate Tool 2 to compare SOMI levels.
Persona SOMI dashboard · three trajectories
Compare the shared city modal split for all personas, Tool 2 persona behavioural trajectory and Tool 3 sentiment-adjusted trajectory. Feedback days are marked on the graph; unsampled dates retain the Tool 2 value.
Without persona behaviour (shared city modal split)With behavioural impact (Tool 2)With behavioural + sentiment impact (Tool 3)
Calculate Tool 2 to see the daily trajectories.
The behavioural difference is Tool 2's policy versus paired frozen modal trajectory. Sentiment adjusts normalized KPIs on reviewed query dates only. The curves show scenario outcomes, not causal estimates of citizen opinion.
Advanced inputs · import JSON, review comments, change sensitivity
1 · GO-DATA response
Paste a post object or a one-post array containing comments with id, body, sentiment and score, plus tracked_at or last_analyzed_at. The score is treated as classification confidence pending confirmation from GO-DATA. The attached city field is literally "string"; confirm that the post belongs to Antwerp. The date can be corrected below.
2 · Gemini classification
Use the prompt with your existing Google Cloud Python/Gemini service. Import its JSON output, then review mappings, citizenship and possible sarcasm. No API credentials are stored in this page.
3 · Comment review
Unmapped comments, noncitizen content, irrelevant comments and flagged sentiment remain excluded. Review each selected KPI and sentiment label. A person named in a comment is not automatically a persona.
Import comments to begin.
Open full comment audit and edit selections
| Use / ID | Comment and quality review | Sentiment | Pillar / KPI | Persona | Relevance |
|---|
Apply reviewed feedback
Apply reviewed comments to the chosen days. Sensitivity controls the size of each normalized KPI change. The dashboard updates automatically.
A
Current runValidation register
B
Provenance & assumptions
C