DOIBoost Dataset Dump

Research in information science and scholarly communication strongly relies on the availability of openly accessible datasets of metadata and, where possible, their relative payloads. To this end, CrossRef plays a pivotal role by providing free access to its entire metadata collection, and allowing other initiatives to link and enrich its information. Therefore, a number of key pieces of information result scattered across diverse datasets and resources freely available online. As a result of this fragmentation, researchers in this domain end up struggling with daily integration problems producing a plethora of ad-hoc datasets, therefore incurring in a waste of time, resources, and infringing open science best practices. 

The latest DOIBoost release is a metadata collection that enriches CrossRef (October 2019 release: 108,048,986 publication records) with inputs from Microsoft Academic Graph (October 2019 release: 76,171,072 publication records), ORCID (October 2019 release: 12,642,131 publication records), and Unpaywall (August 2019 release: 26,589,869 publication records) for the purpose of supporting high-quality and robust research experiments. As a result of DOIBoost, CrossRef records have been "boosted" as follows:

• 47,254,618 CrossRef records have been enriched with an abstract from MAG;
• 33,279,428 CrossRef records have been enriched with an affiliation from MAG and/or ORCID;
• 509,588 CrossRef records have been enriched with an ORCID identifier from ORCID.

This entry consists of two files: doiboost_dump-2019-11-27.tar (contains a set of partXYZ.gz files, each one containing the JSON files relative to the enriched CrossRef records), a schemaAndSample.zip, and termsOfUse.doc (contains details on the terms of use of DOIBoost).

Note that this records comes with two relationships to other results of this experiment: 

1. link to the data paper: for more information on how the dataset is (and can be) generated;
2. link to the software: to repeat the experiment