MY SIMULATOR

Bike Network
Planner

Protected lanes, safe routes, more cyclists.

INTERACTIVE 3D

Bike share Car delay 0Risk 0FPS 0

mixed trafficpainted laneprotected laneclick a street to upgrade it

BIKE SHARE

RISK INDEX1.3

CAR DELAY0.1 min

BUDGET LEFT●3.00 M

Bike Network Planner: Protected Lanes, Safety-Weighted Routes & Mode Share

Click streets to turn mixed traffic into painted or protected bike lanes. Cyclists re-route by perceived stress, commuters re-choose between bike, car and bus, and congestion settles to an equilibrium. Spend a fixed budget well.

Click tool

Network result

Bike share vs money spent

Missions

Independent numerical checks

Also run outside the browser: node src/check.js (Dijkstra vs Floyd–Warshall, flow conservation, logit sums, MSA gap, monotonic planner, BPR curve).

Illustrative static model, not real planning advice. All parameters marked [assumption] (stress multipliers, risks, capacities, costs, logit coefficients) are model values chosen to be plausible, not measured data. The 3D city is drawn at a compressed scale (streets 80 units = 400 m in the model).

Controls

Click or tap a street in the 3D view to change its lane type (tool selector above); drag to orbit, wheel / pinch to zoom; Hero / Junction / Map buttons (keys H, J, M) move the camera.
Auto-plan spends the remaining budget on the protected lanes with the best gain per cost. Reset clears all upgrades.
Heat-map colour: green = few cyclists, yellow = medium, red = busy. The cyan graph shows junction nodes and street edges the model uses.

The science

Graph and stress-weighted shortest path. Junctions are nodes, street segments are edges with length L. A cyclist minimises perceived length c_e = L_e s_k · BPR, where the stress multiplier s_k is 1.9 / 1.35 / 1.0 for mixed traffic / painted / protected [assumption], inspired by level-of-traffic-stress ideas. Paths come from Dijkstra's algorithm.

Mode choice (multinomial logit). P_m = e^U_m / Σ e^U_j with U_bike = −0.06 t + −0.0012 ΔL_stress + 0.1, U_car = −0.045 t + 1.0, U_bus = −0.04 t (t in minutes, ΔL_stress = extra perceived metres) [assumption].

Congestion (BPR). t = t₀ (1 + 0.15 (q/C)⁴) for cars and bikes; a protected lane takes road capacity from cars (800 vs 1200 veh/h/direction) [assumption].

Demand. Gravity model T_ij ∝ P_i A_j / d_ij^1.5, scaled to 22 000 trips/h.

Equilibrium. Method of successive averages (MSA): flows move a fraction 1/n toward the new loading each iteration; 30 iterations on the page, the printed gap is the relative change of the last step.

Risk index. bike-km weighted by relative risk per lane type (1.0 / 0.6 / 0.2) [assumption], increased by car volume. Not modelled: signals, junction geometry, hills, weather, safety-in-numbers, induced demand, e-bikes.

Bike NetworkPlanner

Bike Network
Planner