CVRS in 2010 Report

Perhaps more than I ever wanted to know about the state of the birth and death registration system in South Africa circa 2010: http://unstats.un.org/unsd/demographic/CRVS/Technical%20report%20SADC%20final%20v2.pdf

Capture

Leave a comment

Filed under global health, Uncategorized

Tools They Use

I’m always on the lookout for new tools that I should be using in my daily work. Here is a short write-up by #hackebola researcher Caitlin Rivers on the subject: http://www.caitlinrivers.com/blog/the-setup-tools-i-use-to-track-emerging-infectious-diseases

Leave a comment

Filed under global health

Python Pandas Intros

I’m going to give a Python Pandas guest lecture in the Python Science class next week, and I thought I’d take a look at the Pandas intros that are out there. There are a lot now! Here are some that I flipped through:

http://pandas.pydata.org/pandas-docs/stable/10min.html

http://nbviewer.ipython.org/gist/fonnesbeck/5850375

http://www.gregreda.com/2013/10/26/intro-to-pandas-data-structures/

http://www.gregreda.com/2013/10/26/working-with-pandas-dataframes/

http://www.gregreda.com/2013/10/26/using-pandas-on-the-movielens-dataset/

http://synesthesiam.com/posts/an-introduction-to-pandas.html

http://www.datarobot.com/blog/introduction-to-python-for-statistical-learning/

http://www.kevinsheppard.com/images/0/09/Python_introduction.pdf

http://blog.kaggle.com/2013/01/17/getting-started-with-pandas-predicting-sat-scores-for-new-york-city-schools/

Its fun being a teacher in the age of information.

Leave a comment

Filed under education

Les Roberts blogs from Sierra Leone

A few links to Les Roberts blogging from Sierra Leone:

http://pfmhcolumbia.wordpress.com/2014/10/12/les-says-farewell/

http://pfmhcolumbia.wordpress.com/2014/10/12/oneheavygig/

http://pfmhcolumbia.wordpress.com/2014/10/12/les-roberts-the-prediction-landscape-is-looking-bad/

http://pfmhcolumbia.wordpress.com/2014/10/14/les-roberts-day-10-leaving-columbia-to-be-at-peace-with-columbia/

Leave a comment

Filed under global health

Verbal Autopsy Analysis Slides

I gave a talk on automatic methods to map from verbal autopsy interview results to underlying causes of death last August, and I like the slides so much that I’m going put them online here. Cheers to the new IHME themed templates for Power Point!

Capture

2 Comments

Filed under machine learning

Digging into the GBD 2010 Risk Factors II

A follow-up on my recent adventure digging into the GBD 2010 risk factors, wherein I helped a fellow researcher find the specific diseases to which the risk factor DALYs are attributed (http://ihmeuw.org/2c6v ):

Capture

I helped find where to look at the risk factor attribution, but what if you want to download the numbers? The “download chart data as CSV” link gives you something, but it is not the risk factors, it is just the disease fractions of total DALYs:

Capture

Fortunately, a javascript guru with a few minute told me the incantation to get the risk factor fractions out, recorded here for posterity:

$.fileDownload(app.settings.baseURL+'php/risk_data.php',{ httpMethod: "POST", data:{ measure: app.getMeasure(1, 'percent'), sex: app.getSex(1), age: app.getAge(1), year: app.getYear(1), location: app.getLocation(1), cause: d3.keys(app.metadata.causes), risk: app.getRisk(1), format: 'csv', title: app.charts[1].getTitle(), } });

Leave a comment

Filed under global health

Anyone want to fix things in mpld3?

People are actually using mpld3. It would be great if there was more progress addressing the many issues that this use has uncovered. Interested?

Thanks for your interest in this project. I think that all of these points can be addressed, but it would be helpful to have a minimal example of python code that generates the issue in point (1) reliably. The github issue tracker has discussions related to points (3) and (4), and something that might be related to point (1). For point (2), it would be great to have a specific example in mind also, so that we can be sure any solution reduces file size substantially without compromising graphically accuracy.

  1. https://github.com/jakevdp/mpld3/issues/226 https://github.com/jakevdp/mpld3/issues/250
  2. Would be good to add an issue: https://github.com/jakevdp/mpld3/issues/new
  3. https://github.com/jakevdp/mpld3/issues/247
  4. https://github.com/jakevdp/mpld3/issues/198

As far as when all of these issues will be addressed, that is a pitfall of certain open-source projects that you might already be familiar with from your work with [related project]… I suspect that each fix will require a few hours of debugging at least, with (2) being easiest and (4) being hardest. I have a long list of issues to address, and although I’m happy to put these on it, I never seem to make progress on any of them.

Pull requests are certainly welcome, and if you or your collaborators wants to make these improvements, the mpld3 project will be happy to incorporate them into the codebase.

–Abie

Leave a comment

Filed under software engineering