Science Magazine Data Stories Contest submission

Science Magazine recently hosted a  data stories visualization competition. I cobbled something together but there were some truly excellent submissions http://www.sciencemag.org/projects/data-stories/finalists. I think my favorite was the social network experiment on the campus in Denmark (How People Gather).

Below is my a 1-pager to put my submission (Crowd Intelligence in Disaster Response) into context. 

This visualization is based on Ushahidi mobile phone and internet reports, submitted after the 2010 Haiti Earthquake. The reports were first interpreted. The data was then used to construct ad-hoc digital maps with details (structure integrity, road conditions, hazards, injuries, etc.) relevant to first responders and disaster relief staff in the post-earthquake situation.

Ushahidi was launched in 2008 when a group of scientists were motivated to use their technical expertise to assist with post-election violence occurring in Kenya. Any new technology often comes with new policy implications and crisis mapping is no exception. The following blog post by Patrick Meier, director of crisis mapping at Ushahidi (2009-2012) provides some of his final thoughts on some policy issues that emerged with the technology in the context of the Haiti disaster (https://www.ushahidi.com/blog/2010/04/15/crisis-mapping-haiti-some-final-reflections).

After seeing the data points come to life on the screen one might start wondering if first responders and disaster relief teams might even make use of the archived data long after a disaster and eventually gain insights to better understand for example how communication spreads after a disaster and if they might be able to identify patterns in the data (using time and content of the reports themselves) that would allow predictive modeling to help them develop strategies or procedures that would make them more effective in future disasters. This NSF video short provides an example of one team of researchers at the University of Delaware's Disaster Research Center who are working in this area (https://www.youtube.com/watch?v=uNkQZivo-Ag).

This type of crisis mapping work has continued to serve as a tool in everything from Snowmageddon to the Deep Water Horizon Disaster, to the Arab Spring, to Swine flu. Similar types of data are even provided by transportation systems such as Washington DC's capital bikeshare program. Scientists working with this data could help optimize transportation systems, focusing on the way in which kiosks are restocked with bikes, predicting future demand for bicycles, monitoring maintenance requests, or the data may be used to gain a better understanding how people migrate across a city (https://www.capitalbikeshare.com/system-data). There are other examples of how related technologies are being applied toward similar ends. Geospatial technologies are currently used to document cultural destruction that occurs as a result of conflict such as that occurring in Syria and Iraq (http://www.aaas.org/news/human-rights-day-highlights-scientific-contributions-prevent-cultural-heritage-destruction). 

As our appliances, phones, and transportation systems become more sophisticated more opportunities for leveraging data will arise (as will the policy implications). The work of scientists is never done as they work endlessly to identify and leverage this emerging data while addressing the accompanying policy implications in a responsible and ethical manner. 

The data used in this video is freely available from Ushahidi
(https://datahub.io/dataset/ushahidi/resource/81d058a8-173a-49d9-8ce9-4edf5e7cafc9).

The Role of Code-Archiving Projects in the Peer Review Process

Today the software and code that drive many scientific studies (and ultimately their results) are far from standardized. Often seen as a means to an end this code may not be made available to readers of scientific journals or even the reviewers. Can projects like Open ABM affect positive change when it comes to the peer review process; making it easier to verify that code often hidden down in the weeds of a scientific study?



Scientists working with computational models are typically reporting on the validity of a particular model when they present their results (i.e. how well it fits the data and its usefulness as a predictor for the phenomenon under consideration). But what happens when there is a moth in the relay, a bug in the code? A recent Nature news article by Erika Check Hayden addresses the verification issue and some of its implications while highlighting a Mozilla project aimed at putting such code through a review process similar to that associated with commercial software products. Although it appeared in the competition's science magazine, I am obligated to advance science and serve society so here it is.

Mozilla plan seeks to debug scientific code
Software experiment raises prospect of extra peer review.
Erika Check Hayden
24 September 2013

Note that in the comments section on the Nature site, the author points to Open ABM as a model for getting source code out into the world.

"Erika Check • 2013-09-25 08:16 PM
These didn't fit in the story, but two other projects trying to encourage peer review of code are CoMSES Net (http://www.openabm.org/) in ecology and social sciences and..."

--Tim Kochanski

Some Inspiration for Stranded Research Programs

Or maybe just the inspiration to say... well it could be worse. It is not uncommon for a research program to experience all sorts of delays, setbacks, and even disasters before the midpoint of the funding period. I suspect many reviewers of program progress reports would actually say these obstacles are under reported, which is unfortunate since explaining how one survived the experience is potentially of great value to future investigators. That said, many if not most teams find a hack, work around, or fix that allows them to proceed and produce outputs close to what was originally envisioned and proposed to the funding agency.

I just heard a +Science Friday story on the Kepler satellite that broke down (~4 years after launch), over 40 years aways by rocket and traveling at over 2,000 mph. Technically the project ran to full term but as Ira suggests in the interview, (you've got to be a bit peeved that your relatively new hardware broke so soon with so much unrealized potential). Given that the original research program cannot carry on as usual collecting data, the team is now taking proposals for other potential missions for the stranded satellite. Ideas?... Stories of heroic salvage?... Here is the story on Science Friday.

AUG. 23, 2013

A Telescope Fails, but the Hunt for Exoplanets Continues

This artist's concept shows the Kepler spacecraft. Image credit: NASA/Ames/JPL-Caltech

Click  to enlarge images

The ailing Kepler planet-hunting telescope cannot be fixed, the victim of failed reaction wheels required to aim the instrument. However, researchers still have reams of data to sift through. William Borucki of NASA and Joshua Winn of MIT discuss the search for distant planets.

Presidential Hopefuls Get a Good Googleing

Google Insights for Search is a neat little app that allows you to get a better sense of what Google users are searching for by region and over time.  It was made famous with Google Flu by demonstrating that by just using Google searches for words like "flu" we could see how the flu was spreading across the nation and do so about two weeks before the Center for Disease Control predictions allowed.

Below is the current 3-month report for Google searches conducted nationwide which included the terms "Obama" or "Romney".  If you would like to learn more about Google Insights for Search, use the tool, or just keep checking in as November approaches you can do so by Googling it.

Crisis Mapping with Netlogo

This is a mapping of Ushahidi mobile phone and internet reports submitted after the Haiti Earthquake.  To create the simulation a Google earth map was imported into Netlogo with geo-tagged reports from Ushahidi.  The data is freely available from Ushahidi here: http://community.ushahidi.com/research/.  This sort of mapping has already been done with the Haiti data but I just wanted to see if this could be done with Netlogo. In addition to the geo-locations and timing of the reports it would also be interesting to conduct a content analysis of the text contained in the reports.

 

This type of analysis could prove useful in identifying patterns in how crisis data are reported, ultimately helping first-responders in future disasters.  Some other potential applications for this type of analysis might include bicycle traffic flows in Washington DC's based upon the Capital Bikeshare program data, which is also provided free to researchers and can be downloaded here: http://www.capitalbikeshare.com/system-data.  If anyone can think of any other geo-tagged data sources that might make for interesting simulations it would be great to hear about them.  I would also be interested in hearing of other analyses applied to similar data sets e.g. networks analysis, data mining, pattern recognition, etc.  It would be especially interesting to find data associated with the OpenGov movement www.data.gov so that researchers can help governments improve service levels associated with local, state, and federal governments.