OS: Win 10,
Gramps: 5.1.6
Graph view is showing very complex tree.
Is there any way to generate full tree but in more readable way.
OS: Win 10,
Gramps: 5.1.6
Graph view is showing very complex tree.
Large trees with pedigree collapse are always complex because of the overlapping lines which can cross large parts of the tree. Usually you only show a subset of the family tree for better readability.
I am assuming you are using the All Connected option for the view. This option adds the in-laws of the in-laws for the active person.
It is always a balance. How many generations to display (both up and down) and how wide each generation becomes using All Connected.
The subject of visualizing a complex tree has come up intermittently in the various support forums.
A lot of different strategies have been suggested. Among others, suggestions have included various add-ons, custom filtering, and use of external visualization tools. These overview strategies have been discussed so often that we should have more than enough “grist for the mill” to write a good wiki article. An article would allow the strategies to be presented in an understandable way.
However, just collecting the discussions will take a fair amount of effort… and the 5.2 preparations have begun to consume all free-time of volunteers. Perhaps there is someone who is willing to do the basic research for an article? That is searching the various forums and compiling a list of links to the postings.
(If you are also to trying write an article, that would be great! But if you don’t want to take on learning to format the wiki’s typographical conventions with MediaWiki markdown, feel free to compose in your favorite word processor. Other experienced WikiContributors can volunteer to convert the formatting for you.)
If anyone remembers a favorite response about how to visualize a complex tree, please link it as a reply. (You don’t have to commit to the big research project!)
You can try a a plotter using roll paper. You’ll be limited by the length of the roll. But the critical point, already mentioned, is the number of edge intersections and the angle between intersecting edges (if too small you can’t tell where the edges emerge).
Personally, I no longer use this solution because it proves unreadable as soon as your chart has more than ~200 persons. I prefer to split huge trees into small targeted pieces, e.g. descendant tree over at most 5 generations or ascendant tree up to 7-8 generations, with “connection points” clearly identified to link easily with other partial trees because these trees are not random and they can be arranged easily to minimise still possible intersections.
Actually, if you need any other type of graphs, not in Gramps, you can save a graph report of your choice to Graphviz format and open it in latest version of Gephi.
In Gephi you can also import multiple files to the same graph project.
So, a workflow can be:
I have not exported anything from Gramps for over a year, nearly not used Gramps the last 2, so I do not know if there is any new graph reports that contain both people, events and places as separate nodes…
When I tried to get some more advanced graphs for analyzing huge networks of people, ships, journeys and other events and link them to sources with citations as edges, I found that I had to export multiple different formats/reports from Gramps and export a graph format from Zotero and merge all of them in Gephi/Cytoscape/Tulip with lots of editing, to get a useable graph for my purpose.
You will need to test out a little regarding LABELS, because when showing labels in Gephi or Cytoscape they can often be really messy, lots of overlapping etc.
Do not select too much information for to show in the label, linebreak is not easy if even possible in Gephi
Better to export some report with attributes as a CSV file if needed.
An interactive network graph in Gramps would be great, where you could choose to show people, events, places and sources as nodes, families and repositories as “clusters” containing their respective people and sources and using all relations between people, links to events, links to places and citations as Edges.
And using the Place hierarchy as sub-graphs.
In addition, Notes could have been linked as Nodes, using the Note Type as Edge.
Of course, there had to be a selector for what data you actually wanted to be displayed, and the normal Gramps filter system for data presented…
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.