Back in the early days of FAIRsharing we added graphs which showed the relations between records collected by our “collection” record type. These proved to be popular, and so part of our plan for redesigning the entire site was to expand this feature to cover all records.
But, there was a bit of a snag. FAIRsharing has expanded considerably since then, and we have also added a more complex means of recording relations between records, to better represent the real life relationships between the resources we catalogue. Both of these things make it more difficult to plot graphs which render quickly in a browser and are easy to see.
We continually work on improving features as time allows, and we’ve been working on a new version of these graphs which will allow us to show a lot more of the connectivity between resources. To do this, we’ve made use of new software.
Firstly, Neo4j. This is a graph database specialising in tracking relations between records, and so would seem ideal for our purposes. In order to follow these relations from a starting point we’ve also used APOC, a plugin which offers the ability to follow links to and from a record rather than just in one direction. This is particularly important as our new relations system records the direction of the relation, e.g. record A recommends record B.
The second new component is Sigma.js. This is a Javascript graph plotting library capable of plotting large numbers of points quickly using your computer’s graphics processor. It’s not easy to customise the appearance of plots, but it makes up for this shortcoming by its speed.
You can take a look at an example here.
FAIRsharing continues to evolve and this is not necessarily the last iteration of how we’ll display these data. For the moment, though, we hope you find the new graphs to be a useful and interesting tool.
The image above is a view of the ENCODE Peak information format graph expanded to show three degrees of separation from this record. You can browse this graph at https://fairsharing.org/graph/32.
Instructions on using our new graphs can be found here.
