Colombia 2026 added: result + a new press layer
#2
by AFOS-Analytics1 - opened
The runoff is decided (de la Espriella 49.66% × Cepeda 48.70%, ~1pp), so the Colombia divergence dataset now closes the loop: prediction market × polls checked against the real result.
The honest read: the market and the polls both called the winner correctly but overstated the margin (market priced 88.5% to win, polls a ~8pp lead) for a ~1pp near-tie. Right direction, overstated confidence. The spread is the signal.
New in this update:
- data/colombia-results.csv: the result (official CNE escrutinio pending)
- news/colombia-2026-press-coverage.csv: a structured press-coverage index (22 dated headlines, 15 outlets, ES/EN/PT), the qualitative third axis (market × polls × press). A coverage index, not a sentiment signal.
Dataset (CC BY 4.0 / Apache 2.0): https://huggingface.co/datasets/AFOS-Analytics1/colombia-2026-electoral-divergence
