davanstrien/finepdfs-temporal-stats-fin

该数据集名为网络是否变得更加教育化？，主要对芬兰语（fin_Latn）在105个CommonCrawl转储文件中的教育质量进行了时间序列分析。数据集包含两个主要部分：全局统计（global_stats）和时间统计（temporal_stats）。分析结果显示，从2013年到2025年，高教育内容的比例从4.0%增长到11.6%，平均教育评分从1.55增长到1.76。数据集处理了211,328份文档，共计2,604,770,921个标记，平均教育评分为1.720，高教育内容比例为10.2%。

The dataset is titled Is the Web Getting More Educational? and focuses on the temporal analysis of educational quality in Finnish (fin_Latn) across 105 CommonCrawl dumps. It includes two main components: global_stats and temporal_stats. The analysis shows that from 2013 to 2025, the rate of high educational content increased from 4.0% to 11.6%, and the average educational score rose from 1.55 to 1.76. The dataset processed 211,328 documents with a total of 2,604,770,921 tokens, achieving an average educational score of 1.720 and a high educational content rate of 10.2%.