Road RiskR.I.S.K. research platform

Roadway Infrastructure Safety Kinematics

SciFest research projectApp V2 LiveRSA Recorded Collisions OverlayEvidence Package Exports

Modelled road-safety analysis for Ireland's roads

Road Risk combines public road geometry, controlled scenario assumptions, local model outputs, and historical RSA recorded collision context to compare road sections and build transparent evidence packages.

Open App V2Judge RouteReview MethodView Results

Model outputs are comparative research indicators, not official road-safety ratings, legal guidance, or observed crash predictions.

Private DashboardLegacy App

Public Review

Start with the Judge Route, then open App V2

The public site now has one clear review route before the live workspace. This keeps judging and technical review separate from private account features.

Page Purpose

What to read first

For a first review, read the purpose and boundary first, then move through Method, Results, and the live app.

First read
Use Judge Route for the five-minute review sequence, evidence boundary, and demo handoff.
Primary action
Open App V2 when you want to select a road, inspect local bins, and export evidence.
Public review
Project, Method, Results, Authorship, and References can be reviewed without signing in.
Boundary
Outputs are comparative model results, not official safety ratings or crash predictions.

What Road Risk Does

A live workspace for comparing road geometry, model outputs, and context

Road Geometry Analysis

Road Risk starts with public road geometry, road class, curvature, radius, segment length, and available map tags.

Scenario and Vehicle Controls

Weather, visibility, surface, lighting, behaviour, and vehicle assumptions are explicit inputs, not hidden claims.

Local Risk Bins and Graphs

Selected roads are broken into readable local bins, then compared through distributions and percentile views.

Observed Collision Context

RSA records are used as historical context for comparison, not as proof that the Road Risk model is correct.

Recommended Review Path

Read the project in five steps

For a judge, reviewer, or first-time technical reader, this is the shortest route through the public site before using the live app.

01

Judge Route

Use the five-minute route for the question, hypothesis, variables, evidence boundary, demo, and authorship links.

02

Project

Start with the research question, scope, review boundary, and short demo path.

03

Method

Read how public data, derived geometry, scenario assumptions, and outputs are separated.

04

Results

Inspect reviewed public model-output cases and the public/private evidence boundary.

05

App V2

Open the live workspace to select a road, compare context, and export evidence.

Public pages and App V2 can be reviewed without signing in. Account and Dashboard are used when a user wants to save private projects, imports, datasets, and evidence records.

How It Works

From selected road to reviewable evidence

01
Select a Road

Choose a public road geometry and inspect the available source context.

02
Calculate Modelled Outputs

Apply explicit geometry, physics, scenario, and confidence assumptions.

03
Inspect Local Bins

Find where the selected road has stronger or weaker modelled output.

04
Compare Observed Context

Use RSA recorded collisions as historical context without treating them as model proof.

05
Export Evidence

Save JSON, graphs, Maths traces, and evidence packages for review.

App V2 Capabilities

Inspect roads, compare context, and export the reasoning

The current live app is the primary Road Risk workspace. It keeps the model, map, RSA context, graphs, imports, exports, and evidence surfaces close together.

Selected Road Analysis

Pick a road, inspect public geometry, and read local-bin model outputs with confidence notes.

Route Analysis

Compare modelled route samples and identify the sections carrying the strongest local output.

Hotspot Overlay

Sample nearby road geometry around a chosen place and review local peaks without calling it an official map.

Isochrones

Explore reachable road context around a center point and compare modelled samples inside travel bands.

RSA Recorded Collisions

View historical RSA recorded collision context through a server-backed overlay, not a public CSV dump.

Graphs and Maths

Open distribution views, percentiles, and Maths & Method traces to see how outputs were assembled.

Import / Export Centre

Save, replay, import, and export Road Risk JSON records for reviewable project evidence.

Private Dashboard

Signed-in users can organise projects, imported records, datasets, and Evidence & Reports.

Observed Collision Context

Historical RSA records are context, not model proof

RSA Recorded Collisions
A server-backed observed-data layer

App V2 can request filtered RSA recorded collision points from a server-side source. Fatal, serious injury, and non-serious injury markers help users compare historical context with the selected model output.

This layer is not Road Risk model output and is not official validation of the Road Risk model.

Collision Context

RSA

Historical recorded collision records, not Road Risk outputs.

Selected Radius

25-100 m

Compare nearby records using explicit distance settings.

Comparison

Bins

Selected-road bins can be compared with nearby observed records.

Dashboard and Evidence Workflow

Keep App V2 outputs organised in a private workspace

Private projects

Signed-in users can create projects and organise App V2 exports, imports, datasets, and evidence records.

Source manifests

Dashboard records keep display names, source labels, model/schema versions, hashes, loadability, notes, and tags.

Evidence & Reports

Evidence records can be named, reviewed, opened in App V2 where loadable, and used as report-ready material.

System Scale

Current public dataset and platform surface

Mean output

0.5946%

Reviewed public cases only.

Highest output

4.5096%

Highest stored reviewed case.

Longest stopping

130.9 m

Longest stored stopping distance.

Lowest safe speed

7.9 km/h

Model estimate, not a speed limit.

Vehicle profiles

12

Vehicle presets in App V2 that affect assumptions rather than labels alone.

Graph views

9

Distribution and percentile views for selected-road and sampled-network interpretation.

Export formats

5

JSON, evidence packages, graph data, Maths traces, and route/overlay exports in App V2.

Results Preview

Featured reviewed cases

These are reviewed public model-output cases from the deployed static dataset. They are selected mechanically from stored values, not invented examples.

Highest outputLow

R422

Secondary / Typical

Annualised output
4.5096%
Safe speed
26.0 km/h
Stopping distance
69.3 m

Reviewed public model-output case. Not an official road-safety rating or crash prediction.

View in Results
Lowest outputLow

Unnamed Selected Road

Service / Typical

Annualised output
0.0065%
Safe speed
30.0 km/h
Stopping distance
17.5 m

Reviewed public model-output case. Not an official road-safety rating or crash prediction.

View in Results
Lowest safe speedLow

Unnamed Selected Road

Service / Typical

Safe speed
7.9 km/h
Safe speed
7.9 km/h
Stopping distance
31.8 m

Reviewed public model-output case. Not an official road-safety rating or crash prediction.

View in Results
Longest stoppingHigh

M4

Motorway / Typical

Stopping distance
130.9 m
Safe speed
49.9 km/h
Stopping distance
130.9 m

Reviewed public model-output case. Not an official road-safety rating or crash prediction.

View in Results

Model Pipeline

A transparent chain from road data to saved records

01

Open road data

OpenStreetMap-derived ways, nodes, and tags provide public road geometry and context where available.

02

Selected road geometry

The app analyses a selected segment rather than treating an entire road uniformly.

03

Curvature and radius

Coordinate sequences are converted into heading change, radius, curvature, and bend context.

04

Physics checks

Safe speed, slip ratio, centripetal demand, reaction distance, and braking distance are estimated.

05

Scenario assumptions

Rain, fog, fatigue, overspeed, BAC, vehicle type, lighting, surface quality, and traffic proxy are explicit selected assumptions.

06

Comparative context

Outputs are interpreted using sampled distributions and percentiles under the same assumptions.

07

Reviewed case evidence

Case-evidence JSON is exported explicitly by the user. Public committed cases require review before they become part of the Results dataset.

Scientific Boundaries

What Road Risk does not claim

Responsible interpretation

Road Risk supports comparison, explanation, scenario sensitivity, and evidence preparation. It does not replace collision statistics, engineering audits, field inspection, or professional road-safety assessment.

Road Risk is not an official safety rating.

Road Risk is not a legal speed recommendation.

Road Risk is not an observed crash prediction.

RSA records are historical context only and do not prove causation.

Next Step

Open the live workspace or read the method

Primary

Run App V2

Select roads, test assumptions, inspect bins, use RSA context, and export evidence.

Open App V2
Method

Understand the model

Read the data pipeline, equations, assumptions, and interpretation boundary.

Read Method
Results

Inspect public cases

Filter, chart, and inspect the reviewed public model-output cases.

View Results
Secondary

Legacy App

Open the preserved original app for comparison and historical continuity.

Open Legacy App