Text and data mining FAQs

For information about TDM and how it works please see our text and data mining policy.

To mine full-text content hosted on ScienceDirect you will need to use our API to download content which is specialized for text mining purposes. You can access the API via our developers portal, which will require you to self-register before being automatically sent a personalized API key. Registering for an API key is a simple process that takes a matter of minutes, after which you will be able to download the material you want to mine and have access to our technical support and assistance should you need it.

APIs are used to ensure our platforms can provide a stable and efficient service for all users. TDM typically involves the bulk downloading of vast amounts of content. If this were to occur on the ScienceDirect platform rather than via an API, it could disrupt or delay the system's performance and/or stability. This disruption could negatively affect researchers visiting the platform simply to read and/or download articles for non-TDM related research.

We need to ensure that we can support the simultaneous needs of millions of human readers and dozens of text miners. In order to serve all use cases efficiently it makes sense to separate the traffic into different channels optimized for each use case.

Self-registration is the process by which you register for an API key. During this process, users are reminded of the terms and conditions of using the API key. An additional provision relates to users who wish to use the API key for text and data mining purposes.

No, there are no hard limits on the number of items that may be downloaded via our API. Nevertheless, a reasonable and customary rate limit remains in place to ensure equal access to the API for all users, and we continue to ask users to use our service responsibly.

We understand the need to be flexible and continue to monitor usage and consult with researchers. However, we reserve the right to deactivate any API key if we believe usage is abusive or impacting the stability of our systems.

A TDM corpus is a collection of downloaded material accessed via Elsevier’s API. This will typically include copyrighted material from books and journals, open access articles and supplementary materials.

After receiving the API key, you are able to download a corpus to text and data mine it.

Please note, our TDM services are not designed to facilitate sharing of individual articles, datasets or any other inputs, and restrictions apply to sharing your TDM corpus. We have specific guidelines on how we facilitate sharing of articles and research data.

No. Your API key and an extracted corpus is personal to you and should not be shared with third parties, including within your institution's repository or within social collaboration networks. This is because our TDM service is not designed to facilitate sharing of individual articles, datasets or any other inputs. We do support scholarly sharing, but this is separate to our TDM service, and further details on this can be found here.

Instead, you can share a reference to the contents of a corpus by creating a list of DOIs of the documents contained in your dataset. This will enable any researcher with access to the underlying content contained in the corpus to recreate it by retrieving the same set of DOIs using their own API key.

No. However, you can post a reference to the contents of a corpus by creating a list of DOIs of the documents contained in the dataset. DOIs are resolvable and will provide transparency as to the sources you have used.

Yes, you can combine an Elsevier TDM corpus with material from other publishers. We recommend doing this by using the CrossRef API available via CrossRef Text and Data Mining.

Our TDM service is designed to support non-commercial text and data mining for academic research purposes and facilitates access to a TDM corpus for this purpose.

Neither a corpus nor its individual inputs should be used for other purposes. Our agreement therefore sets out a number of ways in which a TDM corpus cannot be used.

This includes using a TDM corpus or any of its individual inputs for commercial purposes:

• Users may not, for example, make a profit from publicly posting the DOIs of the articles contained within a corpus, either in its entirety or in part.

• Indirect commercial activity can include, for example, associating advertising with a freely-posted corpus or delivery to third parties.

Users are also not permitted to use the corpus in such a way that could:

• Compromise existing Elsevier products and services.

• Substitute for existing Elsevier products and services.

• Replicate existing Elsevier products and services.

For example, publicly posting a TDM corpus for free access is prohibited, as this could have such effects. This restriction applies regardless of whether the user profits financially from the posting.

A TDM output refers to the results of text and data mining. This output may include both a researcher’s extracted results, alongside snippets from a corpus to provide context.

When distributing your TDM output externally, you can include a few lines of text of individual full-text articles or book chapters to provide context to your results.

You can do this by including snippets of up to 200 characters surrounding and excluding the text entity matched, or by including bibliographic metadata. Where snippets and/or bibliographic metadata are distributed, they should be accompanied by a DOI link that points back to the individual full-text article or book chapter.

A notice is also required in the following form: "Some rights reserved. This work permits academic research purposes only, distribution, and reproduction in any medium, provided the original author and source are credited."

We recognize that placing your results in context is helpful, and the limit of 200 characters has proven sufficient for the majority of researchers currently using our TDM services. However, please contact us via our Permission Helpdesk if you require more extensive quotations as we are always happy to assist.

You do. We do not claim copyright over your TDM output (i.e., your extracted results).

Yes. There are no restrictions on where and how you can publish your TDM output. The conditions we place on reuse relate to the original copyright material (the TDM corpus) you have used to perform TDM, which is subject to the original copyright restrictions of the articles within.