useNews

The useNews dataset has been compiled to enable the study of online news engagement. It relies on the MediaCloud and CrowdTangle APIs as well as on data from the Reuters Digital News Report. The entire dataset builds on data from 2019 and 2020 as well as a total of 12 countries. It is free to use (subject to citing/referencing it).

The data originates from both the 2019 and the 2020 Reuters Digital News Report (http://www.digitalnewsreport.org/), media content from MediaCloud (https://mediacloud.org/) for 2019 and 2020 from all news outlets that have been used most frequently in the respective year according to the survey data, and engagement metrics for all available news-article URLs through CrowdTangle (https://www.crowdtangle.com/).

To start using the data, a total of eight data objects exist, namely one each for 2019 and 2020 for the survey, news-article meta information, news-article DFM's, and engagement metrics. To make your life easy, we've provided several packaged download options:

- survey data for 2019, 2020, or both (also available in CSV format)
- news-article meta data for 2019, 2020, or both (also available in CSV format)
- news-article DFM's for 2019, 2020, or both
- engagement data for 2019, 2020, or both (also available in CSV format)
- all of 2019 or 2020

Also, if you are working with R, we have prepared a simple file to automatically download all necessary data (~1.5 GByte) at once: https://osf.io/fxmgq/

Note that all .rds files are .xz-compressed, which shouldn't bother you when you are in R. You can import all the .rds files through `variable_name <- readRDS('filename.rds')`, .RData (also .xz-compressed) can be imported by simply using `load('filename.RData')` which will load several already named objects into your R environment. To import data through other programming languages, we also provide all data in respective CSV files. These files are rather large, however, which is why we have also .xz-compressed them. DFM's, unfortunately, are not available as CSV's due to their sparsity and size.

Find out more about the data variables and dig into plenty of examples in the useNews-examples workbook: https://osf.io/snuk2/

《useNews》数据集之编纂旨在助力对网络新闻互动之研究。该数据集依托于MediaCloud与CrowdTangle API，并整合了路透社数字新闻报告（http://www.digitalnewsreport.org/）之数据。整个数据集以2019年及2020年的数据为基础，覆盖总计12个国家。数据集供公众免费使用（但需注明出处或引用来源）。

数据来源包括2019年及2020年的路透社数字新闻报告（http://www.digitalnewsreport.org/）、MediaCloud（https://mediacloud.org/）所提供的2019年及2020年媒体内容，以及通过CrowdTangle（https://www.crowdtangle.com/）获取的所有可用新闻文章URL的互动指标。

为便于数据使用，共提供八个数据对象，分别为2019年及2020年的调查数据、新闻文章元信息、新闻文章DFM（Document Feature Matrix）以及互动指标。为了简化操作流程，我们提供了多种打包下载选项：

- 2019年、2020年或两年的调查数据（亦提供CSV格式）
- 2019年、2020年或两年的新闻文章元数据（亦提供CSV格式）
- 2019年、2020年或两年的新闻文章DFM
- 2019年、2020年或两年的互动数据（亦提供CSV格式）
- 2019年或2020年全年的数据

若您使用R语言，我们还准备了一个简单的文件，可一键下载所有必要数据（约1.5GByte）：https://osf.io/fxmgq/

请注意，所有.rds文件均采用.xz压缩格式，在R语言环境中使用时不会给您带来困扰。您可以通过`variable_name <- readRDS('filename.rds')`导入所有.rds文件，而.RData（亦为.xz压缩）文件则可通过`load('filename.RData')`简单导入，这将把多个已命名的对象加载到R环境中。对于其他编程语言，我们也提供了相应的CSV文件。然而，这些文件体积较大，因此我们也进行了.xz压缩。遗憾的是，由于DFM的稀疏性和体积问题，无法提供CSV格式的DFM。

更多关于数据变量信息及丰富的示例，请参考useNews-examples工作簿：https://osf.io/snuk2/