Welcome to my Open Notebook

This is an Open Notebook with Selected Content - Delayed. All content is licenced with CC-BY. Find out more Here.

ONS-SCD.png

Research Protocol For Manitoba Centre For Health Policy

This post has moved here: http://ivanhanigan.github.io/2015/10/research-protocol-for-manitoba-centre-for-health-policy/.

Version control

2013-12-02: This post was originally released
2015-10-02: The URL to the University Manitoba guidelines changed and has been updated.

Posted in  disentangle Project Management


graphviz-automagic-flowcharts

  • Back in 2009 Joseph Guillaume worked with me on a complicated workflow
  • He came up with this python script to help keep track of the steps
  • The simple text file is a list of transformations, inputs and output
  • It is converted to the right format and graphviz creates a html page with pop-outs

Simple text list

Transformation
        description
                the thing 
        inputs
                the thing before
                another thing
        output
                the next thing
        notes
                the thing is this other thing   this is a really long description 
                blah blah asdfasdfasdfasdfasdfa 
 
Transformation
        description
                yet another thing
        inputs
                the next thing
        output
                a final thing
        notes
                this is a note

  • keep this in your work directory and update it whenever you add a step to the workflow
  • the list can be as big as you like (hundreds of steps), and entered in any order, the inputs/output relationships determine how the graph looks at the end
  • to run the script just do the one line

Python code: run

python transformations.py workflow_steps.txt index

  • open the html page and click on a square box to bring up the pop-out
  • short text is shown, long text is replaced by an ellipse and only shown in pop-out

Conclusions

Posted in  research methods


Gantt Charts For Health Professionals

Gantter

  • This is an online tool, that can integrate with Google Drive http://www.gantter.com/
  • It also imports MS Project files

Taskjuggler

gantt-tj3.png

Posted in  research methods


sharing-and-extending-research-protocols

  • Last week I started a new discussion about sharing and extending research protocols.
a site for sharing research protocols. The idea is to give people
a place to post research protocols since everyone develops them
and then mentions them in papers but they rarely make it online in
a usable format. We see this as a big bottleneck in the rate of
knowledge discovery. Since users will always want to change and
adapt a given protocol for changing applications or new technology
or just to improve it. This sort of thing is done with open source
all the time (i.e. git) but how you'd do it with an online
protocol based out of a wiki is unknown. The current version of
the website is not very flexible and doesn't really enable the
sort of dynamic collaboration that I envision would make this a
really useful tool. Git seems like an obvious conceptual starting
point but there'd need to be a front end markup system in place
that was integrated with that git-type backend. Since you've
mentioned git a few times and the open science project, I thought
you might find this interesting to think about. any suggestions or
inspirations would be welcomed

My past experience

  • I have previously tried to use a wiki for the development of new sections in an existing study protocol on estimating the effects of bushfire smoke air pollution and human health.
  • We combined elements of two major air pollution/health study protocols (one from Europe, one from America) with local considerations and decisions made within our team.
  • On reflection all this was mostly useful for me, to a lesser extent the other programmer and I don’t think the project manager or statistician ever even looked at any of this.

Jeff Leek’s current approach using GitHub

I have been keeping an eye on the work by Jeff Leek on putting protocols onto Github and inviting collaboration to edit and extend them as he says:

I put it on Github because I'm still not 100% sure I got it
right... I would welcome feedback/pull requests on how we can
improve the policy to make it better

Leek Group policies and protocols on github so far

I forked “datasharing” for my own. Renamed: “How to share data to avoid misunderstanding”

  • I liked the datasharing policy so much I forked it for my own collaborations.
  • I initially just made a minor recommendation to one of the lines in the original using the great Github feature that if you edit a repo you don’t have access to, it forks the repo and creates a feature branch behind the scenes. After you submit the changes (which is a commit) it puts you right into the pull request form.
  • Therefore it can all be done on the GitHub site. I submitted the pull request without leaving my web browser.
  • But then I thought I would have done things a bit differently. In particular I’d like a webpage with a table of contents, and a nicer looking landing page.
  • so I cloned my fork of the repo, then created a new branch called gh-pages, then copied the original to a new file called index.org, added some Emacs Orgmode HTML export magic and then “C-c C-e h” and WHAMM-O I’ve got my own version for extending with at http://ivanhanigan.github.io/datasharing/.

The code version:

git clone git@github.com:ivanhanigan/datasharing.git ~/tools/datasharing
cd ~/tools/datasharing
git checkout -b gh-pages
mv README.md index.org
touch README.md
# edits to index.org, "C-c C-e h" and WHAMM-O
git add index.org index.html README.md
git push origin gh-pages

  • In this way I think people can share and collaborate easily (if they know git)
  • as well as easily take work started by someone else and extend it for their own work.
  • I wonder a little bit about how much of my extensions I should offer back to Jeff Leek as the original author.
  • I guess he can always track what I do to it via the link back to his original shown in https://github.com/jtleek/datasharing/network

PS Jeff Leek uses github a lot!

As he says in this post from the future-of-statistics unconference:

You can read it on Github here (https://github.com/jtleek/futureofstats).
I put it on Github for two reasons:
 
- I agree with Hadley's statement that the future of statistics is on Github.
- I summarized them based on my interpretation and would love
  collaboration on the document. If you want to add your new
  thoughts/summaries, add a new section with your bullet pointed
  ideas and send me a pull request!

Conclusions

  • Jeff Leek is on to something really interesting with these github policies and protocols
  • I find the look of his plain markdown README.md pages a bit un-inspiring
  • but then I spend too much of my time tweaking my html by far (while Jeff is getting on with the real work of publishing papers)

Posted in  research methods


our-paper-is-out-heat-health-warning-systems

Finally our article has been published in final version in Environmental Health.

  • Title: The impact of heat on mortality and morbidity in the Greater Metropolitan Sydney Region: a case crossover analysis
  • Authors: Wilson Ann Leigh, Gerard Morgan Geoffrey, Hanigan Charles Ivan, Johnston H Fay, Abu-Rayya Hisham, Broome Richard, Gaskin Clive, Jalaludin Bin,
  • Environmental Health.2013, 12:98.
  • DOI: 10.1186/10.1186/1476-069X-12-98
  • URL: http://www.ehjournal.net/content/12/1/98

Finally……. thanks to everyone, especially Geoff Morgan, Richard Broome, Fay Johnston; Bin Jalaludin

And especially Leigh Wilson.

Posted in  extreme weather events