Speaker: Christine Stohn, bX project manager
There are two components — the recommender and hot articles.
This began in 2009 with the article recommended, and as of this year, it’s used by over 1100 institutions. This year they added the hot articles service, with “popularity reports”. And, there is a mobile app for the hot articles service. Behind the scenes, there is the bX Data Lab, where they run experiments and quality control. They’re also interested in data mining researchers who might want to take the data and use it for their own work.
The data for bX comes from SFX users who actively contribute the data from user clicks at their institutions. It’s content-neutral, coming from many institutions.
bX is attempting to add some serendipity to searches that by definition require some knowledge of what you are looking for. When you find something from your searching, the bX recommender will find other relevant articles for you, based on what other people have used in the past. The hot articles component will list the most used articles from the last month that are on the same topic as your search result.
It currently works only with articles, but they are collecting data on ebooks that may eventually lead to the ability to recommend them as well.
The hot articles component is based on HILCCÂ subjects that have been assigned to journal titles, so it’s not as precise as the recommender.
You can choose to limit the recommendations to only your holdings, but that limits the discovery. You can have indicators that show whether the item is available locally or not.
It’s available in SFX, Primo, Scopus, and the Science Direct platform. Hot articles can be embedded in LibGuides.
Atmetrics – probably will be incorporated to enhance the recommender service.
They are looking at article metrics calculated as a percentile rank per topic, which is more relevant today than the citations that may come five years down the road. It’s based on usage through SFX and bX, but not direct links or DOI links.
