C
Cycling City
1BM130 · TU/e
Buurten—
Critical—
High—
Ø Access—
Ø Usage—
Ø Bike—
Ø Audit—
Spatial view
Neighbourhoods —
Buurt ↑ Gem. Access ↑ Usage ↑ Pred. ↑ Bike ↑ Audit ↑ Priority
—
Selected

—

—
—
Access
—
—
Usage
—
—
Bike
—
—
Audit
—
—
Gap
—
—
In muni
peer set
Cycling share per neighbourhood — COROP 36 CBS · consumptie
≥ 30% cycling
20 – 30%
10 – 19,6%
5 – 10%
< 5%
ODiN 2024 — travel mode COROP 36 · n ≈ 6.700
ODiN 2024 — Travel Mode Eindhoven (COROP 36)
This dashboard shows the results of a machine-learning analysis on the Daily Travel Survey (ODiN 2024) dataset, filtered for COROP region 36 (Southeast North Brabant). An XGBoost classification model predicts the transport mode based on trip characteristics and socio-demographic variables. Analysis estimated on n ≈ 6,700 trips.
Modal share — COROP 36 Eindhoven (ODiN 2024)
🚗
46,7%
Car
n ≈ 3.130
🚲
19,6%
Bicycle
n ≈ 1.310
🚶
18,7%
Walking
n ≈ 1.250
⚡
9,1%
E-bike
n ≈ 610
🚌
5,9%
OV
n ≈ 390
Modal share — horizontal distribution
🚗 Car
46,7% n≈3.130
🚲 Bicycle
19,6% n≈1.310
🚶 Walking
18,7% n≈1.250
⚡ E-bike
9,1% n≈610
🚌 Public Transport
5,9% n≈390

Source: ODiN 2024 · COROP 36 (Southeast North Brabant) · Trips (Verpl=1) · Hvm ∈ {1,2,7,8,9}

XGBoost model
81,2%
Accuracy
0,81
Weighted F1
2.016
Test set
10
Features
XGBoost multi:softprob SMOTENC oversampling 70/30 train/test stratified split

5 classes: Car · Public Transport · E-bike · Bicycle · Walking

Feature importance (SHAP — relatief)
Travel distanceAfstV
1,00
Transit travel timeReisduurOP
0,79
Household car ownershipHHAuto_DANS24
0,83
Age classKLeeft
0,71
Disposable incomeHHGestInkG
0,54
Destination typeDoel
0,47
Travel purposeKMotiefV
0,38
Day of weekWeekdag
0,33
Number of tripsAantRit
0,23
Week numberWeek
0,19

Relative SHAP mean |φ| values — normalised on AfstV. Indicative based on feature permutation analysis.

Accuracy per travel purpose
🏢 To/from work 84,9%
💼 Business/professional 85,9%
🩺 Services/personal care 91,6%
🛒 Shopping/groceries 81,3%
🎓 Education/course 81,2%
👥 Visiting/staying over 81,3%
🎉 Social/recreational 78,0%
❓ Other purpose 86,1%

Separate XGBoost models per travel purpose. Touring/walking: insufficient data.

Cycling share per municipality — COROP 36 vs. ODiN average (19.6%)

Source: CBS Neighbourhood Map 2024 (bike_share_consumption per neighbourhood, weighted average per municipality) · Dotted line = ODiN 2024 COROP average (19.6%) · Only municipalities with ≥ 3 neighbourhoods in the data.

Classification report — Full model (n_test = 2,016)
Transport mode Precision Recall F1-score Support
🚗 Car 0,840,920,881.057
🚌 Public transport (train/bus) 0,840,810,8370
⚡ E-bike 0,810,480,60190
🚲 Fiets 0,740,690,72369
🚶 Lopen 0,780,780,78330
Weighted average 0,810,810,812.016

Model: XGBoost multi:softprob · Dataset: ODiN 2024 · Region: COROP 36 (Eindhoven) · Features: AfstV, AantRit, ReisduurOP, Doel, Weekdag, KLeeft, Week, HHAuto_DANS24, HHGestInkG, KMotiefV

Selected neighbourhood

—

—
—
Cycling
—
—
Car
—
Access
Municipality Rankings
Amenity & Cycling Analysis
Neighbourhood Explorer
x = bike access advantage  |  y = composite score
CBS Buurtkaart 2024
OSM amenities · ODiN 2024
Model GBM-v3.1 · R² 0.74
Showing — buurten