Omolumeter - JohnTigue/idots GitHub Wiki

Omolumeter (pronounced OMO·lum·eter) is the main end user experience of this project. It is liberally licensed open source, the code for which is in this repository. The Omolumeter code is deployed at omolumeter.com, but can very easily be incorporated into other web sites as all that is required is to host some static files. Please do so and feel free to reach out for help. Contact John Tigue via email to discuss this.

In techno-babble:

Omolumeter is the proof-of-concept (read: reference implementation of the Spec), front-end, Web browser-based, white-label widgetry (read: single page apps, optionally iframe embeddable) that can parse and visualize outbreak data found on the Web.

Description

With tongue in cheek, outbreak_time_series can be described as modern open-source anti-virus software. That phrase can be broken out as follows.

Modern: this means highly interactive web browser based visualizations leveraging SVG, via D3.js. outbreak_time_series's dependencies works with any browser newer than IE8 (which is already fading fast in late 2014, at less than 10% of global usages, and Microsoft will stop supporting it in 2016). outbreak_time_series also uses AnjularJS 1.3.x, another project which recognizes that IE8 is not worth the limitations it imposes on any forward looking effort. So, this is designed to be useful for the foreseeable future.

Open-source: The code is Apache 2.0 licensed (that is, do with it freely as you wish, commercially or otherwise). Data for the 2014 Ebola Outbreak in West Africa is also made available, as free open data (PDDL licensed). We have curated data and provide it via a RESTful web service API based on the Outbreak Time Series Specification.

Anti-virus software: Ebola is a virus and this software project is anti-ebola (and pro-geeky humor).

Status

The Omolumeter as deployed at omolumeter.com has the following visualization:

The above are the most basic visualizations. The next visualizations to be added are:

  • Static map
  • Animated time map

Future development

After the above, the Spec (in-development) will have been exercised enough in term of reading from Spec-compliant documents. That will inform the spec. E.g. does the JSON translate into an elegant JavaScript object model (read: JavaScript API), etc.

The other dimension that the Spec needs to be exercised is in mapping to existing CSVs via CSVW schema. And surely it will turn out that various CSVs will need to be proxied on omolumeter.com, probably with some data translation (read: limitations of Spec, or simply things the Spec should not do simply because there is a legacy of such having happened). Or, even just having CORS set up such that omolumter.com can be an API server to arbitrary web browser clients with Spec visualizers running in pages hosted on any website on the web.

Finally, after all the above, it will be time to make the existing Omolumeter visualization much more sophisticated data analysis and explanation machinery, specific to this one task: get more eyeballs seeing outbreaks.

For example, the Epicurves view should have a config dialog for selecting which countries to plot; currently that is hardwired to the top three highest deaths countries. Speaking of hardwired, only by-country data is viewable. Clearly this should be expanded to states, counties, city, and other arbitrary administrative units or even custom regions.

Currently only one data source is plotting on the epicurves. It would be very interesting to plot numbers simultaneously from multiple sources. This is the sort of thing that a standard data interchange format makes much easier than is currently possible on the web.