In our previous post, we discussed Edward Tufte’s notions about the rich visual expression of data, and we introduced some basic metaphors used in genealogical diagramming: trees, fans, roads, timelines, quilts, and networks. We observed any one of these can express perhaps one or two classes of data very well, but often at the expense of other kinds of data. Our conclusion: to achieve Tuftian richness, we’d need to mix our metaphors.
This is another way of saying that, in most instances, I’m not entirely satisfied with the balance between simplicity and salience found in the visualizations produced by most genealogy software. This reminds me of the dissatisfaction I had with standard numerical ordering systems in genealogy. In response, I invented a hybrid of a Meurgey de Tupigny System, a Register System, and the version numbering convention used in software development. I figure if I could propose a new numbering system, why shouldn’t we try looking at genealogical data visualization in an equally new way?
To arrive at such a mixing of metaphors, we first need to go deeper into understanding them. We’ll look at various examples from the history of art and datavis to help with our synthesis.
Roots and branches
The tree metaphor, with its roots and branches, is the most recognized diagram metaphor used in ancestry and descendancy charts. Despite its aptness, a tree can get unwieldy as a database gets quite large, and the metaphor breaks down as families inevitably marry into a line, inducing a kind of vine-in-the-tree or tree-into-tree sub-metaphor. The “husband and wife” tree illustrated below is an example of a gemel tree (produced through insoculation), and suggests something close to this reality, but the grafting has to be multiplied again and again in a manner that soon becomes unsustainable.
The metaphor of roots and branches emerges when we consider a person exists both as an ancestor and a progenitor, something you surprisingly don’t see often enough in genealogy charting. In a discussion with Mark Humphrys, the Atlantic’s Steve Olson comes to this realization:
… in essence every one of us who has children and whose line does not go extinct is suspended at the center of an immense genetic hourglass. Just as we are descended from most of the people alive on the planet a few thousand years ago, several thousand years hence each of us will be an ancestor of the entire human race—or of no one at all.
Steve Olson, The Royal We, the Atlantic, May 2002
1 | Centered on my great-grandfather Theodore James Cromar, my tree in Ancestry exhibits the “hourglass” phenomenon discussed by Steve Olson and Mark Humphrys above. | Screen-capture by the author from Ancestry
2 | A Celtic Tree of Life symbol | Pebbles
Tree of Life
A Tree of Life or World Tree symbolism exists in innumerable cultures, from ancient Egypt to the Aztecs, from Judaism to Norse mythology. A well-known version of this symbol is the Crann Bethadh, the Celtic Tree of Life. With its elaborate stylization and balance between depictions of both root and branch, it becomes an apt metaphor for the human experience, with each one of us being at once a descendant and a future ancestor.
Fans and mandalas
When a tree chart gets too unwieldy, a fan metaphor is introduced as a space-saving device. Essentially, it does this by bending a linear order into a radial one. Compare these two diagrams of my portion of the tree at Family Search, both of which were generated at TreeSeek.
A tree diagram versus a fan chart using the same dataset: my portion of the family tree at Family Search.
The relationship between these diagrams is roughly analogous to that found in Earth map projections. The relationships at the far end of the tree, representing the first few generations, tend to “flatten out” like the poles on a Mercator projection map, making relationships harder to understand. The fan chart tends to display those relationships “at the pole” with compactness, clarity, and less visual distortion, making comparisons easier, rather like the conic projection for a hemisphere.
Mercator maps project the Earth’s surface onto a rectangular area, significantly distorting the poles. A conic projection limited to a hemisphere significantly reduces distortion: compare the area of Greenland to the rest of North America. | All map images from PROJ
Mandala: a full fan
To take even more advantage of a radial arrangement, one can change the metaphor from a fan to a mandala. There is a natural argument in favor of a full radial chart over an ordinary pedigree: the circumference increases with each successive generation. As the number of nodes grows, so does the available space to render them. This quality of a radial chart is initially quite appealing, but as we add successive layers we realize the linear growth of the circumference is no match for the exponential pace of generational growth.
1 | M. C. Escher, Circle Limit III, 1959. As implied by the title of this mandalic print, the physical circle has a limited capacity to carry the fractally infinite pattern of growth. | Wikimedia Commons, Fair Use
2 | My dataset from the Family Search tree showing 10 generations. Beyond 10 the diagram becomes unwieldy. | TreeSeek
Rivers and roads
In 2021, Stuart Sly captured this dramatic image of three famous Scottish bridges. They span not only the Firth of Forth but also three centuries of engineering progress. | STV News
Rivers and roads share some hierarchic characteristics which they can lend metaphorically to connections between people, routes of migration, and the passage of time. These can be considered master metaphors, since structurally they underpin other metaphors. Time works like an arrow — one-directional — and dimensionally sequential in our experience, so rivers are more like timelines, whereas roads can be both sequential (think subway line or bus route) and multi-directional (think city blocks). Roads can inform the structure of quilt and network diagrams.
A sub-genre of the river metaphor is the waterfall diagram. This presents a descendancy in the manner of heading outlines in a manuscript. My great-grandfather Theodore James Cromar yields a waterfall that contains dozens of people over 5 generations. A waterfall containing the known descendants of 10 generations will carry thousands and be quite unwieldy.
A very long diagram needs a more compact expression, and the long scroll strip map design of John Ogilby’s beautiful Britannia road atlas is iconic in the field of information graphic design (Ogilby was born, as it happens, in Scotland).
Timelines and progress bars
People have struggled with the visual expression of time since time immemorial. Joseph Priestly, an English polymath who was granted a Doctor of Law at the University of Edinburgh, was responsible for developing time datavis expressions in the late 1760s, creating conventions that are still in use today. He was influenced by the work of Jacques Barbeu Dubourg, who created one of the first instances of expressing time along an x-axis in an unbroken, uniformly scaled manner. His Chronographie timeline is over 50 feet long and housed in a special scroll box. We take timelines as so natural an expression of their function as to be self-evident, but these guys were inventing them!
Read more about these and other influential visualizations and their inventors at info we trust… and while at it, visit their incredible Interactive Timeline of the Most Iconic Infographics. You’ll love the Internet again.
Quilts and trellises
Quilts in the textile world and trellises in the architectural world rely on weaving or intersecting elements to generate a pattern of relationships. It’s hard to imaging how this kind of visualization could be informative in genealogy, but the inventors of Geneaquilts at Aviz combined rows representing individuals with columns that represent family groups to express genealogical datasets containing several thousand individuals (like mine at 4000 and counting!). Their project brief describes the interesting methodology and use of interactivity to navigate the weave of relationships generated in a Geneaquilt.
It’s not a standalone, however, or at least one that I have access to on my MacBook Pro. Instead, it has been incorporated into other genealogy software like GRAMPS and Progeny. In Progeny, it has been rebranded as a Trellis Chart.
The Greek Pantheon displayed in a Geneaquilt chart. The big outliers that form rectangles indicate the incestuous relationships that Hades and Zeus become involved in. | AVIZ
Nodes and networks
Along with trees, another metaphor that can express causal relationships are network diagrams, also known as Lombardi diagrams. These were named in honor of Mark Lombardi, a Neo-Conceptual visual artist whose work asked uncomfortable questions about the abuse of power by financiers and politicians. He himself called these diagram-infused artworks Narrative Structures — were he still alive, one can only imagine the bizarrely baroque narrative structures Lombardi would have created based on Trumpian shenanigans! Such a feat has actually been attempted here.
It helps to understand that nodes in a network can be connected in patterns with varying degrees of hierarchy: centralized, decentralized, or distributed.
Network hierarchies
Observing these 2D diagrams, we are inevitably drawn to the possibility of rendering them in a higher dimension. As Mark Humphrys observed:
The problem is that genealogies aren’t two-dimensional, so any attempt to put them on paper is more or less doomed from the start. They aren’t three-dimensional, either, or you could make a structure. They have hundreds of dimensions.
Mark Humphrys quoted inThe Royal We, the Atlantic, May 2002
Well, that’s a bit over the top, but Mark does have a point. We need higher dimensions to see these relationships, and a 3D structure that is navigable (bringing in the fourth dimension of time) and hyperlinked (a fifth dimension?) would begin to solve a huge limitation. Nodes and networks are hardwired to do this kind of work.
Multi-dimensional trees
Sadly, so-called “3D” family tree offerings in the current market are woefully not up to the task. Most are little more than simple extrusions, layered stacks, or low relief versions of 2D trees, with no new information generated for the trouble. In fact, most of them create less clarity, courtesy of the horrifying graphics.
Progeny Genealogy offers Progeny 3D Family Tree, claiming to be “the only program that can display your family tree in 3 dimensions” — if you can call stacking flat graphic elements so close together that you can only see them at a radically oblique angle a “display.” Meanwhile, a different 3D Family Tree is abandonware whose time has come — and clearly gone. This rendering of the Kennedy lineage is illegible. Finally, Clanview offers a weak VR experience as an add-on to free level trial access in a clumsy, tiered freemium business model. The result is actually less rich in information than a traditional 2D chart. Why would I ever possibly want to see my family masquerading as the International Space Station in low Earth orbit?
It is evident that much needs to be done in the arena of 3D and VR interactive diagramming.
Mixing the metaphors
… and we can probably start hacking away at that problem by mixing metaphors, now that we’ve explored a few. Two existing experiments out there show some real promise.
BloodLines
BloodLines by Michael Plunkett combines a decentralized node/network metaphor with a timeline. I placed a now out-of-date GEDCOM file in the front page of the BloodLines website just to test it out, and it returned a beautifully interactive version of the data instantaneously. In the screen-capture below, the family lines of Cromar (in orange) and Robb (in yellow) dominate the families that marry into the line or children by fathers outside of the line. Try it yourself with your own GEDCOM file and be amazed! You can create a standalone, permanent version by grabbing it at GitHub, but you need a server and some coding savvy. There is some real potential for VR integration with this!
LifeSpans
We’ve praised Courtney Barr’s Lifespan Tree in a prior post. She combines a mandala-style chart with progress-bar style timelines
Upon closer scrutiny of this admirable synthesis, I find myself wishing for greater legibility of generational levels, which generational drift can often obscure. I also note the potential for some confusion in the radial bars that signify parent-child relationships: they are based on parental death, not the child’s birth — and, they are the same line weight as the timeline bar, an unforced error in my opinion.
If we applied this synthesis to a descendancy chart, rather than an ancestry chart such we see above, what might happen? Would we have something entirely new? Could we create a parent-child relationship based more logically on the child’s birth intersecting the parental line, thinning up that line weight issue in the process?
Applying this to my family database
Laborious and tedious as the compilation of this work has been… a variety of views were continually opening upon me during the execution of it…
Joseph Priestley, A Description of a Chart of Biography | A New Edition, with Improvements, p. 22, 1800
These mixed metaphors represent two potential expressions of data for the descendancy chart of Peter Cromar. As of this writing, I took a break about a third of the way through processing Generation 8, when I passed the 4000 person mark. I’ve certainly underestimated the awesome force of nature that is the Mormon obsession with family records! I now anticipate Cromarville will be occupied by about 5000 citizens at the end of the compilation of these records, which I now have to focus on. You may not see another blog post until I slog through the last 20% of my 6th and 7th cousins, every one of them, like me, a multi-great-grandchild of Peter Cromar.
Leave a Reply