r7dai/cinerai-1800

Cinerai-1800是一个多模态数据集，旨在用于电影成功预测和剧本分析。它直接将电影剧本的创意内容与其财务表现指标联系起来。数据集包含1,471个处理过的CSV文件，每个文件代表一部电影，结合了完整的剧本文本（逐场景）与财务元数据和成功标签。数据集的结构是CSV文件集合，每行代表剧本的一个场景或片段，但电影的元数据（预算、收入、ROI）对整个电影是重复的。数据字段包括场景文本、IMDb ID、预算、收入、ROI、成功标签、投票平均和流行度分数。Cinerai-1800 is a multimodal dataset designed for Movie Success Prediction and Script Analysis. It links the creative content of movie scripts directly to their financial performance metrics. The dataset contains 1,471 processed CSV files, where each file represents a single movie. Each file combines the full script text (scene-by-scene) with financial metadata and success labels. The dataset is organized as a collection of CSV files. Each row in a CSV file represents a scene or segment of the script, but the metadata (budget, revenue, ROI) is repeated for the entire movie. Data fields include scene_text, imdb_id, budget, revenue, roi, success_label, vote_average, and popularity.