Public research surface
NewsLens Research Dashboard
Explore sentiment-model evaluation and AI-assisted news analysis through a public, read-only dashboard.
Workspace Index
Core Workspaces
Four entry points, each optimized for a different kind of analysis.
ModelsModel EvaluationCompare sentiment model performance across corpora and inspect benchmark tradeoffs.SandboxText AnalysisRun focused sentiment checks on short passages without leaving the dashboard.MethodAboutReview the project method, data pipeline, and research boundaries.WriteupPoster and PaperUse a publication-ready outline for the poster, paper, figures, and claim boundaries.NewsNews AnalyticsExplore source, topic, lens, workflow, and reliability diagnostics.
Research Focus
Exploratory, auditable, and bounded
NewsLens is designed for exploratory analysis. It supports transparent comparisons of model behavior, source-level framing patterns, topic-controlled differences, and data quality signals without treating automated scores as final judgments.
ComparativeTrack how outlets, lenses, tags, and events differ under shared measurement rules.
TransparentKeep derived metrics visible so users can audit what supports each visual claim.
ControlledFavor topic, tag, and same-event views over pooled summaries when interpretation matters.