tweet-reflect

Recalibrating Twitter content strategy involves analyzing engagement logs to compute per-content-type scores for median likes, replies, and reposts. The system normalizes these scores to 1.0 to update strategy weights and promote top tweets.

Content-type weight normalization computes engagement scores from 30-day performance logs and adjusts strategy weights to sum to 1.0. This mechanism prioritizes high-engagement content types while preserving existing strategy prose.

Promoting top tweets and pruning underperforming accounts requires evaluating engagement scores from performance and tweet logs. High-scoring tweets append to a promoted candidates file, while zero-engagement or stale accounts are removed.

When calculating engagement scores, content types with fewer than three samples are marked as insufficient_data. This constraint prevents inaccurate weight normalization by excluding under-sampled content types from the strategy recalibration process.

Automating strategy updates from jsonl performance logs computes engagement metrics to rewrite strategy markdown files. The workflow appends daily reflection logs and generates promoted candidate records while enforcing data integrity guardrails.