We’ve been working to try and surface relevant information in the Vasari^k Artist Network application – released a few weeks ago – which displays relationships between artists, patrons, family members and friends throughout history. The application stores the data and relationships of around 200,000 people, with varying degrees of detail. Some artists are very well accounted for, others less so, and it will be an ongoing task to fill out the information in a sensible and accurate manner.

Another task is to gain insight from the data available, as otherwise we’re left with a spider web of people and relationships which only makes sense to the most dedicated art historians, and isn’t a particularly useful tool for anyone else. Ideally we want to cater for those users who are just browsing the network, as well as those who want to dive a bit deeper into researching a particular individual and what might have made them tick. To achieve both aims, we’re working on having a more flexible ‘lower’ level, and a more opinionated higher level of queries that you can pose to the network, which joins up all the lower level features to show interesting patterns.

To really draw out information and insight from this spider’s web, we need to work with the information that we currently have, and what tools we could use to display the different perspectives.

• Birth year/ death year
• Nationalities (though not where they spent their time)
• Associated Names
• Degree (e.g. number of relationships to other people, and who was related to them)
• (Some) Citations, though none available on the relationship itself.

We already use degree in our display to change the size of the nodes – larger nodes reflect a more connected individual, but can start to look a bit confused with larger networks.

What if want to find another facet on the data, where perhaps an individual links not just across communities but across generations as well? Fortunately there are layout engines that we can use to help surface that information, and we can base our layout on certain attributes of the person, e.g. birth date, to give us a concentric layout:

Here we can see, for example, that Vasari was key in that respect as he had links from earlier generations (reflected in the middle concentric circle, as well as links to his generation (reflected in the outer circle).

There are additional ways of representing the data and it starts to become clear with the levers that we have available (e.g. layout, colour and size) we can start to represent many more dimensions in the graph and surface important information that isn’t easily available in other formats.

Dead or Alive?

By highlighting different variables in the network we can also start to surface information that is hiding in plain sight; whether individuals close to the individual in question were alive or dead at the same point in time. We can identify that Giorgio Vasari I (Vasari’s uncle) never met his nephew, having died many years before his nephew was born. It also helps to identify bad information in the network (e.g. that a person was a colleague of someone that they had never had the possibility to meet, or that we don’t have access to the individual’s date of death, or the dates of birth and death are wildly incorrect).

Simplification

Sometimes we’re just after who preceded or succeeded an individual, so we’ve added some extra functionality to just pull out the successors or predecessors up to a certain level, and display these in an appropriately hierarchical (or DAG – directed acyclic graph) manner.

Please try the graph out for yourself (it’s free to sign up) at https://vasarik.app. We would be very happy to get any and all feedback!