Newsflo Integration: FAQsNewsflo Integration: FAQs

Newsflo provides the ability to measure an academic's research impact by covering relevant mentions of their work across tens of thousands of mass media outlets around the world. Its core components are:

• Tracking of media mentions for researchers, universities and institutes 
• Sophisticated, high precision search based on Scopus author ID + affiliation ID or Funding body ID.
• Clever clustering of similar stories from different sources - An advanced text clustering algorithm which clusters relevant news articles together within a time based window. Various algorithms help identify if the media mention is about topics such as sports, academic related, crime related, obituaries, but only academic related are imported into Pure.

Near real time feed of news articles from the Nexis Metabse API - coverage from over 45,000 (English-speaking) news outlets in over 20 countries including the USA, India, China, Brazil and all major European countries

Newsflo is a data processing system (enhanced by machine learning) which searches for any author that has been captured in Scopus together with their known affiliations (current and past). This means it does a 'string' search for "author name + affiliation" in any news it receives, e.g. "David Thomson + Pure University" or "D. Thomson + Pure University". As this data strongly depends on Scopus data it is therefore important the data for a given researcher is correctly registered.

A mismatch means we found a media mention for a research professional, which is actually not about that person. This mostly happens for the reasons below:

• The person has a common name which is shared with multiple people, e.g. David Neal
• The affiliation captured in Scopus is very generic. Unfortunately, a lot of mismatches occur because of incomplete or 'noisy' affiliation data linked to an author in Scopus, e.g. "Hospital", "Oxford", "The University"
• Agglomeration of media mentions from news source. Some news sources provide us with input that merge several media mentions in a single file. Our algorithm cannot distinguish where one mention starts and another stops and thus finds the name and affiliation in two unrelated mentions.

The Scopus AU-ID is used to match the names of the researchers in Pure with their names in Scopus and once that match is done, it searches in all Newsflo sources by the author's name.

API users have the ability to disclaim media mentions per author via our so-called 'feedback loop'. This flags the match so it doesn’t occur again for that author, it disassociates the author from a given media mention; allowing the mention to capture other individuals but no longer the one sending the feedback. The feedback should only be availble for each individual author mentions due to GDPR regulation. Disclaiming media mentions: 

• Only available on Press/Media records synchronized from Newsflo
• Cannot be undone; once a Disclaim has been committed, the Person is permanently removed from the Press/Media record in Pure
• If one of the internal persons in the content, disclaim the content, the entire record would be retired
• It is always available, regardless of the workflow state of the record
• After confirming the user wants to continue with the Disclaim, the record will still need to be saved in order to commit the change

The record would be removed from Pure.

 To prevent media mentions from being imported you can use the option to block persons from the Newsflo association list. This will ensure that no Newsflo content will be created or linked with the person.

News media is a fast moving information sharing field and news get out of data rapidly. For this reason most media sources do not sustain their hyperlinks for more than 3 months. Therefore also hyperlinks provided via Newsflo (via LexisNexis) can become outdated.

1. Scopus data changes
2. Media sources can vary with time
3. New features. E.g. name variants

• Online News: Web based news publication
• Print: clippings or text that was originally made available in print
• Blogs
• Comments: comments which are posted on blogs or some news articles

Pure's data model does not allow to change templates from one content type(press/media) to another (Research output) or perform an automatic migration between them, so the right way to proceed is to disclaim the clipping and import the record from the correct content type and source you consider more accurate.

This depends on the set up of the workflow steps for the imported records coming from Newsflo. 

If you select to have the imported records in the workflow step 'Entry in Progress',  personal users will be able to edit part of the content in the media coverage records, but if you set up the workflow step for the imported records in the workflow step 'Approved',  only  users with editorial roles and rights for this content type will be able to edit some fields on the imported records.

The Newsflo product is continuously working on developing the matching algorithm it uses to import more accurate clippings. Please note that we currently use the disclaim feature as a "delete" strategy, the difference is that the feedback is sent to the product when you disclaim for them to consider in their improvements.

No, notifications won't be sent when these records are disclaimed, but if you encounter this, please report this to Pure-support and we'll help you review the Press/media content type configurations.

E.g., under the user settings, you'll find that the options to enable email notifications on this are not available as it is for other content types like RO.

In case you identify wrongly assigned records coming from NewsfLo:

1. Review if the authors has an incorrect Scopus Author ID assigned to them in their record in Pure. If that is the case, you can correct that ID in Pure and the next time the sync job runs, those records will clean up the wrong Press/Media records. 
2. If the record in Pure has the correct Scopus Author ID:
   1. Use the Feedback-loop or Disclaim the content. 
   2. If the issues with the feed seems to be affecting most of your researchers, submit a case via Product Support Portal including specific examples and if possible the Scopus authorIDs and Scopus affiliationIDs so we can investigate your issue.