Within a workbench, you can create and have access to ten datasets at a time. Each dataset can include up to 2,000,000 documents. To create a dataset, click the create new dataset button. You will have the option to either select publication titles or select ProQuest databases.
Select publication titles
Find content: There are over 40,000 titles to choose from. These publications range in source type and include scholarly journals; trade journals; magazines; historical newspapers; newspapers; books; conference papers and proceedings; blogs, podcasts, and websites; audio and video works; wire feeds; and other sources. In addition to source type, each publication title also includes subjects, date range, and whether you can mine the full text of the publication. Click next when finished.
Refine content: Once you select the publication titles you need, you can refine the content by searching through the documents within the publications by keyword, date published, source type, and document type. Click next when finished.
Create dataset: You will be provided with a summary that includes the number of documents within your corpus as well as the publications from which those documents are drawn from. You will need to provide a dataset name and description, and click create dataset.
Select ProQuest databases
Find content: There are close to 300 databases to choose from. Each database listed includes a synopsis of the database coverage and whether you can mine the full text of the database. Click next when finished.
Refine content: Once you’ve selected the databases you need, you can refine the content by searching through the documents within the databases by keyword, date published, source type, and document type. You can use ProQuest's searchable fields to narrow your search results. Click next when finished.
Create dataset: You will be provided with a summary that includes the number of documents within your corpus as well as the databases from which those documents are from. You will need to provide a dataset name and description, and click create dataset.
A dialog box will appear that says: Submission successful! Your dataset has been successfully submitted. It may take time to process but you can close this message and return to your dashboard to check the status.
You will be redirected to the workbench. Each dataset includes a name, date range, search, data source, document count, status, date created, an option to delete, and an option to export metadata. When your dataset is being created the status will read in progress, but once it has been processed the status will say ready for Jupyter.
To export a dataset's metadata, click the export button. You can export the citation metadata or the extended metadata in a CSV format. Click the metadata dictionary listed under the extended metadata button to learn what constitutes extended metadata. Make sure to read the legal notice and note the weekly export limit of 1,000,000 document metadata records. After selecting citation metadata or extended metadata, click export. The metadata will download to your computer in a zip file.
| University Libraries