Tree Metadata Survey Results & Next Steps (..and Hello!)

[Elliott Hauser]

** Apologies for cross-posting; I think these results are relevant to both MIAPA and OTOL **

Hi all.

I've been helping Karen Cranston analyze the results of the survey that went out earlier this Fall, and am very excited to be involved with Open Tree.  I'm a first-year PhD student in information science at UNC studying data provenance in computational science.  Hello to OTOL and hello again to MIAPA-discuss!  

Our initial findings from the survey are that respondents say that phylogenetic metadata is relatively easy to produce compared to how useful it is.  There are plenty of things we need to investigate further, but this is an interesting and surprising start.  We were expecting 'consumers' of trees to want lots of metadata and 'producers' of trees to say it was onerous to produce it.  But if our results are representative of the population of phylogenetics researchers broadly, that the bottleneck in data quality and archiving is not lack of easy tools but lack of a perception of usefulness of the metadata itself.   Or perhaps even an actual lack of usefulness of the metadata (which I know will sound like heresy to some).

The slides Karen and I gave at NESCent this morning are here:

Phylogenetics & Data Provenance: Survey Results

NESCent will be putting up a video of the talk soon; link to follow.  There was some great discussion afterwards, so I recommend taking a look.

The key graphs are attached for a TL;DR version.  The graphs are stacked one dimensional axes, each representing a metadata category, and ordered by descending importance; there is no y component.  You can should be able to clearly see the divergence in importance and ease of production.  The presentation doesn't have (m)any details of the survey methods on the slides yet; I'm adding those in.  See also planned writeup below.  A commentable version that I'll be editing and improving is here:

https://docs.google.com/presentation/d/1KF4SZ-ikRUUs6A6ZLkzhwC2106nhtlmwrX_AkgCEa90/edit

Please feel free to comment if anything is unclear and I'll add details.  

Feel free to edit, but please be signed in from your Google account when you do so so we know who said what!

Our next step for this will be to write up the results.  We'll also look to investigate further with interviews, another survey, and an analysis of popular phylogenetics software to determine what metatdata outputs they're capable of/default to.  We'll try to describe how the beliefs we've uncovered about metadata utility and ease of production map to real-world behaviors for clues on how to maximize metadata quality and the consequent data re-usability.

Elliott

p.s. you can find me on Mendeley, G+, or Linkedin