Models / mcit

mcit/event-attendance-forecaster

Event Attendance Forecaster

PublicMinimal riskGov ReuseForecastingscikit-learnN/A1.1.0
720 5.2K/30d 38

MAPE (ticketed events)

9.6%

MAPE (free events)

17.8%

MAE (attendees)

1,840

0.84

About this model

Gradient-boosted regressor that forecasts attendance for national events, league fixtures, national-day celebrations, exhibitions, and concerts, across major Qatari venues. Features include ticket-sales pace, competition tier, weather window, school calendar, and historical venue draw.

Intended use

Capacity planning, transport coordination, and staffing decisions for event organisers and government coordination committees.

Training data lineage

Trained on mega-event-attendance-history with demographic context from population-projections for catchment-size features.

Limitations & bias notes

First-of-kind events with no comparable history fall back to venue-tier averages and carry wide intervals. Free-entry community events are harder to forecast than ticketed ones because sales-pace features are absent.

Evaluation metrics

MAPE (ticketed events)9.6%
MAPE (free events)17.8%
MAE (attendees)1,840
0.84
#events#attendance#forecasting#venues#capacity-planning

Try it

Live sandbox

horizon=12

fixture

Feedback

Owning entity

mcMCIT
Updated
2026-01-29
Latest version
1.1.0
License
Government Internal Reuse
Access
Open