HDX Repository - JohnTigue/idots GitHub Wiki
The Humanitarian Data Exchange (HDX) is the place to pay attention to for ebola outbreak data in terms of actual hosted data as well as modern visualizations thereof. See the HDX ebola dashboard for evidence of this statement.
The HDX is a project of the UN with the goal of making "humanitarian data easy to find and use for analysis." The HDX team is not your grandfather's United Nations. Imagine if the Web start-up community has an outpost in the UN: that would be the HDX. They are very much of the open source mentality and they have a lot of code in public GitHub repos. They also do a lot of the group management in public. This is most encouraging.
The project is showing traction:
The HDX site itself is a highly customized instance of CKAN. In the governmental and humanitarian tech crowds, CKAN is a popular "data portal" document repository open source codebase.
There are many datasets on HDX, not all of which are related to the 2014 Ebola Outbreak in West Africa. This page focuses on some of the ebola datasets and code that uses those datasets. This is only a high level overview. For detailed analysis of JSON and API at HDX, see JSON Ebola2014 Data Found on the Web.
This section reports on code from the HDX folks that was found on GitHub. First I mention their API experiment which clearly indicates that they have the big picture in mind (read: Web service API data feeds in JSON).
Since at least 2013-10, these folks have completely understood what needs to be done on the API front, see Data APIs page on their wiki.
In 2014-07, experiments on this front can be seen: https://github.com/luiscape/hdx_api
As of 2014-11-12 they have started experimenting with an API for ebola data. Initially the information is extremely simple, just top line summaries, not the nitty-gritty that is in the Outbreak Time Series Specification. As CJ Hendrix wrote in the announcement:
It’s an experiment!
We have just started to experiment with the CKAN Datastore and with this dataset in particular. Most of our datasets are not datastore-enabled right now. We may shortly be adding sparklines and other visualizations that require a time series of these data, so the structure of this dataset may change
This is great news because time series is part of exactly what this project's Outbreak Time Series Specification addresses. So, John Tigue is interacting with them.
For detailed analysis of ebola related JSON from HDX see JSON Ebola2014 Data Found on the Web.
URL: https://data.hdx.rwlabs.org/ebola
GitHub: https://github.com/OCHA-DAP/hdxviz-ebola-cases-total
This is very important, see JSON Ebola2014 Data Found on the Web for detailed analysis.
2015-01-07: found this:
http://docs.hdx.rwlabs.org/west-africa-ebola-outbreak-visualization/
In 2014-08 Capelo was contributing to an ebola narrative. It's D3.js-based and can run (tested by JFT) via file:// URLs so it should be easily offlined, which is one of the things I am interested in.
https://github.com/luiscape/hdx-datastorer-ebola-cases
Collection of scripts to manage datasets on CKAN's DataStore.
https://github.com/luiscape/hdx-datastorer-ebola-cases/blob/master/data/data.csv
A well data typed CSV file that seems to get updated now and again. New info is appended to the end. It has the top line numbers broken out by country. For each country, info like cases (suspected, confirmed, probable), same for deaths, both of those as Case fatality rate (CFR).
Some code in here fetches data from an URL:
https://data.hdx.rwlabs.org/api/action/resource_show?id=a02903a9-022b-4047-bbb5-45127b591c85
Which points to
https://docs.google.com/spreadsheets/d/1LcGzK41O5xANVxTvympUwBBz_eioQJ7VJqzRh6r5XJc/export?format=csv
Which is a direct download of a CSV file of which contains the exact information which comes out of the HDX top line numbers experimental API. So, I guess the former feeds the latter.
This dataset is the most important one that needs to be paid attention:
URL: https://data.hdx.rwlabs.org/dataset/rowca-ebola-cases
License: public domain
Sub-national time series data on Ebola cases and deaths in Guinea, Liberia, Sierra Leone, Nigeria, Senegal and Mali since March 2014
Ebola outbreak time series data at national and sub national levels since March 2014. Data compiled manually from a number of published reports. Updated by OCHA ROWCA every working day.
This is the first highly usable data I found. The data is available as an Excell spreadsheet, Data Ebola (Public).xlsx. So, they still need to figure out how to get that information easily and automatically available via HTTP but it seems they are working on that.
https://github.com/luiscape/hdx-datastorer-ebola-cases
On 2014-11-14, JFT downloaded this and used it for development purposes.
https://data.hdx.rwlabs.org/dataset/ebola-cases-2014
Total number of probable, confirmed and suspected Ebola cases and deaths in Guinea, Liberia, Sierra Leone, Nigeria, Senegal, Mali, Spain and USA. Extracted from WHO: Ebola Response Roadmap Situation Reports, the latest of which was on 19 November 2014
Related but not outbreak numbers
https://data.hdx.rwlabs.org/dataset/evd-cases-by-district
This data is made available from the WHO - Global Health Observatory. Data on new probable and confirmed cases by epi week and district. There is a file for each country Liberia, Sierra Leone, and Guinea. All the data from the countries can be found at http://apps.who.int/gho/data/node.ebola-sitrep.ebola-country?lang=en There are 2 values for each district. One is from the Ministry of Health SitRep. The other is from the patient database maintained in country. There is latency in getting the patient database is updated therefore the numbers are lower in the latest weeks.
https://data.hdx.rwlabs.org/dataset/ebola-treatment-centers
Ebola Treatment Centers or Units (ETCs or ETUs)
This dataset represents the best-known collection of status and location of the facilities known as Ebola Treatment Centers or Ebola Treatment Units in Guinea, Liberia and Sierra Leone, with relevant attributes and information. Please forward any mistakes or requested changes to [email protected]. Updated frequently, latest update 26 November.