Crate oximedia_analytics 

oximedia-analytics

Media engagement analytics for the OxiMedia Sovereign Media Framework.

Provides viewer session tracking, audience retention curves, A/B testing, multi-armed bandits, cohort analysis, funnel analysis, real-time aggregation, time-series decomposition, EMA trend analysis, geo/device breakdowns, reservoir-sampled heatmaps, CTR tracking, and engagement scoring — pure Rust.

Re-exports

pub use ab_testing::assign_variant;

pub use ab_testing::bayesian_winner;

pub use ab_testing::winning_variant;

pub use ab_testing::winning_variant_with_alpha;

pub use ab_testing::AssignmentMethod;

pub use ab_testing::BayesianAbResult;

pub use ab_testing::Experiment;

pub use ab_testing::ExperimentResults;

pub use ab_testing::OptimisationMetric;

pub use ab_testing::Variant;

pub use ab_testing::VariantMetrics;

pub use bandit::BanditArm;

pub use bandit::BanditStrategy;

pub use bandit::MultiArmedBandit;

pub use bandit::RegretTracker;

pub use cohort::build_cohort_matrix;

pub use cohort::Cohort;

pub use cohort::CohortAnalyzer;

pub use cohort::CohortDefinition;

pub use cohort::CohortMatrix;

pub use cohort::CohortRetentionCell;

pub use cohort::CohortWindow;

pub use cohort::UserEvent;

pub use cohort::ViewerEvent;

pub use engagement::compute_engagement;

pub use engagement::decompose_time_series;

pub use engagement::exponential_moving_average;

pub use engagement::linear_regression_slope;

pub use engagement::ContentEngagementScore;

pub use engagement::ContentRanker;

pub use engagement::DecomposedSeries;

pub use engagement::EmaConfig;

pub use engagement::EmaResult;

pub use engagement::EngagementComponents;

pub use engagement::EngagementTrend;

pub use engagement::EngagementWeights;

pub use engagement::SeasonalPeriod;

pub use engagement::TrendDirection;

pub use error::AnalyticsError;

pub use funnel::compute_funnel;

pub use funnel::compute_loyalty;

pub use funnel::predict_churn;

pub use funnel::ChurnAssessment;

pub use funnel::ChurnConfig;

pub use funnel::ChurnRisk;

pub use funnel::FunnelAnalyzer;

pub use funnel::FunnelDefinition;

pub use funnel::FunnelMilestone;

pub use funnel::FunnelReport;

pub use funnel::FunnelResult;

pub use funnel::FunnelStep;

pub use funnel::FunnelStepDef;

pub use funnel::LoyaltyComponents;

pub use funnel::LoyaltyScore;

pub use funnel::LoyaltyWeights;

pub use funnel::SessionEvent;

pub use geo_device::BreakdownAnalyzer;

pub use geo_device::DeviceType;

pub use geo_device::GeoDeviceReport;

pub use geo_device::PeriodDelta;

pub use geo_device::Region;

pub use geo_device::SessionRecord;

pub use geo_device::SliceComparison;

pub use geo_device::SliceMetrics;

pub use geo_device::TimestampedRecord;

pub use quantile::percentiles;

pub use quantile::TDigest;

pub use realtime::BucketMetrics;

pub use realtime::RealtimeEvent;

pub use realtime::SlidingWindowAggregator;

pub use retention::average_view_duration;

pub use retention::compare_to_benchmark;

pub use retention::compute_retention;

pub use retention::compute_retention_incremental;

pub use retention::compute_segment_retention;

pub use retention::drop_off_points;

pub use retention::re_watch_segments;

pub use retention::ContentSegment;

pub use retention::IncrementalRetentionState;

pub use retention::RetentionBenchmark;

pub use retention::RetentionBucket;

pub use retention::RetentionCurve;

pub use retention::SegmentRetentionResult;

pub use session::analyze_session;

pub use session::analyze_sessions_batch;

pub use session::attention_heatmap;

pub use session::build_playback_map;

pub use session::reservoir_sampled_heatmap;

pub use session::HeatPoint;

pub use session::PlaybackEvent;

pub use session::PlaybackMap;

pub use session::ReservoirHeatmapConfig;

pub use session::SampledHeatmap;

pub use session::SessionMetrics;

pub use session::ViewerSession;

Modules

ab_testing

A/B testing framework for media content experiments.

anomaly

Z-score anomaly detection over a rolling window.

attribution

Watch time attribution: allocate total engagement credit to content segments.

bandit

Multi-armed bandit algorithms for adaptive media experiments.

cohort

Cohort analysis: group viewers by their first-view date and track how each cohort’s engagement evolves over subsequent time windows.

ctr

Click-through rate (CTR) tracking for thumbnails and content previews.

engagement

Engagement scoring model for media content.

error

Error types for the analytics crate.

event_buffer

Generic bounded event buffer with capacity-triggered draining.

fingerprint

Session fingerprinting via FNV-1a hashing.

funnel

Funnel analysis, churn prediction, and viewer loyalty scoring.

geo_device

Geographic and device breakdown analytics for viewer session metrics.

heatmap

Grid-based 2-D heatmap accumulation and density querying.

multivariate

Multivariate testing: test multiple variables (factors) simultaneously.

percentile

Exact percentile computation over pre-sorted data.

quantile

Approximate quantile estimation using a simplified t-digest.

realtime

Real-time analytics aggregation with sliding window metrics.

recommendation

Content recommendation scoring based on engagement similarity.

replay

Session replay reconstruction from PlaybackEvent sequences.

retention

Audience retention curves: compute, analyse, benchmark, and compare.

segment_retention

Segment-level retention analysis for content chapters and sections.

session

Viewer session analytics — playback events, session metrics, playback maps, and attention heatmaps.

uniformity

Chi-squared uniformity testing for variant assignment distributions.

weighted_retention

Weighted retention curves accounting for viewer demographics.