As primeiras datas e aparições da CENA.BR estão chegando. Segue @cumulativeweb no Instagram →
Cumulative Web Inc — logoCENA.BR

Metodologia

O manual que o agente de IA usa para operar a CENA.BR — público, pra qualquer um auditar.

O objetivo

Levar quem procura rap, trap e underground no Brasil até a música da CWI: cada página de cidade é uma porta de entrada que termina em play (Radio 365, YouTube, Spotify).

Rastro de linhagem: R2 — 100 verified Radio 365 listener sessions + 10 verified new signups + 25 unique Neon Nights devices by 2026-10-28.
city pages -> organic search/social discovery -> stream plays on Radio 365 embeds -> measured listener sessions in radio-sessions/summary.json.
Conversão: page visit -> press play on the Radio 365 embed (or follow to Spotify/YouTube) -> session recorded.

A fórmula

Insumos: demand (0-1): search/social demand evidence. scene (0-1): existing artist/event density. gap (0-1): lack of curated competition — higher means more open. confidence (0-1): data quality behind the numbers.

opportunity = round(100 * (0.35*demand + 0.25*scene + 0.40*gap) * (0.6 + 0.4*confidence))

Leitura: Higher gap beats higher demand at the start: with zero domain authority in a market, an open lane outranks a crowded one. Confidence is shown as a separate badge — a high score with low confidence is a hypothesis, not a fact.

Previsão e calibragem: Each city carries predicted_trend + prediction_basis. The weekly loop compares predictions against Search Console clicks and Google Trends movement, then publishes a calibration note. Predictability is earned by calibration, never promised.

O loop semanal

Weekly: re-run scoring with fresh Google Trends (pytrends, $0) + Search Console queries (connected) + page click data; re-rank cities; expand where traction shows; publish calibration note in hidden_files.

Regras de honestidade

Por que cada peça existe