This application combines social media metrics and development activity to create a comprehensive scoring system for crypto assets. It helps identify notable trends and anomalies in both social engagement and development activity.
For each metric and day, each asset is assigned a percentile-like normalized value computed from its rank, so that the top-ranked asset scores 100 and the bottom-ranked asset scores 0, with intermediate assets spread linearly between them. The per-asset social score and dev score are the arithmetic means of these normalized values across the metrics in each category, and they are combined into a composite score weighted 60% social and 40% developer activity. An anomaly signal is emitted when the rolling z-score of this composite score, measured against its own mean and standard deviation over a fixed trailing window, exceeds a configured threshold, and the raw composite score also exceeds a minimum level — ensuring alerts fire only on statistically unusual spikes from assets that are already scoring meaningfully high.