How to make a tree graph
March 17, 2012 12:19 PM   Subscribe

Help me find software to make tree graphs.

Looking for software package suggestions (mac or pc) that will produce something very much like a tree graph, from a table of data points stored in an excel file.

I have several data sets in which a binary “division” event occurs, spawning two new events, and about 100 data points later (typically 1 hour separates each data point), each of these two new daughter events will then also undergo a division event spawning two new daughter events etc. The period between parent events and daughter events is important, and should be reflected on the graph in the distance between the parent and daughter events. I have a large excel spreadsheets of events and daughter events to input but have struggled to find a sufficiently flexible platform for making tree graphs. Any suggestions would be welcomed. Free software would be preferable but I am willing to pay if necessary.

As a bonus, I would also like to display information about a secondary measurement associated with the 100 or so data points of each branch that separate a division event and a daughter event but am a little stuck as to how to visualize this. The Secondary measurements vary in magnitude between each parent and daughter event, typically forming a bell-shaped curve. Simplistically, I though that maybe the secondary measurement for the 100 or so data points could be represented as the thickness of the line connect parent and daughter events but know of no software that could implement this.

Hopefully someone out there can understand this abstract description of my data!

This question has lead me to GraphViz, which I am currently exploring, although I'm not sure that it is sufficiently flexible with regards to the lines linking the nodes.
The tree graphs are intended for use in a presentation, so should look as nice as possible.
Thanks in advance for any help!
posted by SueDenim to Science & Nature (7 answers total) 4 users marked this as a favorite
 
It's mentioned in that previous question already, but I'd second processing. If you're comfortable doing a little programming it's great for creating custom visualisations, and it's very quick to get started (this is an example of drawing a fractal tree).
posted by lucidium at 12:39 PM on March 17, 2012


As far as I can tell, even though your data probably doesn’t have anything to do with evolution, you are basically trying to draw what biologists call a rooted phylogenetic tree. Wikipedia has a list of visualization tools (and links to two further resources) which might be helpful, but I don’t know if any of these programs are general enough to be used for your purpose.
posted by wachhundfisch at 12:39 PM on March 17, 2012 [1 favorite]


This came up last week in an R blog. Good news: R is free and can read your data and make a super slick dendogram with some noodling. Bad news: it may be a lot of noodles. If you have some coding experience, I say go for it.

How to read a dendogram
Making and comparing two dendgrams on the same chart
another example
A central repository of r bloggers
posted by shothotbot at 12:53 PM on March 17, 2012 [1 favorite]


What you're looking at is a dendrogram. R makes dendrograms. Your distance measurements would be used to layout the figure. I'd search around for how-to links, but on preview, just take a look at shothotbot's comment above.
posted by Blazecock Pileon at 12:56 PM on March 17, 2012


I like Graphviz, but I don't think it's the right tool for this job.
posted by hattifattener at 1:46 PM on March 17, 2012


Check out the D3 examples here. Maybe you can use what they call a dendogram.
posted by tayknight at 2:28 PM on March 17, 2012


thanks so far for the answers. the R thing looks like it is the ticket, but my coding skills are basic and my time to expand them is non-existent. maybe if i look around and see if there is a data set and code that i can easily doctor for my own purposes....

D3 looks interesting but it appears that the nodes must all be at the same depth, suggesting that differences in the length of the branches between sister nodes is not tolerated.
posted by SueDenim at 5:10 PM on March 17, 2012


« Older [Apartment Exchange Question] ...   |  I need true and well known exa... Newer »
This thread is closed to new comments.