In some cases, these data sets may come from researchers who have engaged
with clients, cities, regions and governmental organizations and have the ability to then
post (potentially anonymized) versions of the data on their own or other websites.
However, it is desirable to reach out to the agencies, cities, and organizations
themselves to make such anonymized data available.
An example is a city making its traffic data readily available (e.g. as a data feed)
to the public for the sake of encouraging the development of new technologies. In this
way, the traffic data may be used by research scientists for studying impacts of traffic on,
for example, public health or other environmental systems, or the impact of other factors
on traffic. Similarly, a hospital or regional health agency which regroups multiple
hospitals could publish online daily or weekly data on the diseases that they are treating
to allow researchers to develop the models that link those diseases to other systems that
are present in the same geographical area.
We also need to develop standards for sharing data at different scales without
compromising private or proprietary information. One possibility would be a standard for
“scrambling” data in a way that allows models to be first tested on scrambled data.
4.3. Model linking and sharing
While the sharing of data sets may be a precursor to encouraging the entry of
many mathematical researchers into studying the problems related to sustainability,
taking it a step further involves the sharing of mathematical models themselves. The
challenges of studying sustainability are daunting. The ability to communicate openly
about how models work, to use outputs from one model as inputs into another, and to
integrate systems that may have interacting feedback promises to be a valuable, if not
indispensable, goal.
Web platforms where models as well as data can be shared have the potential to
foster valuable dialogue, and to lower substantially the barriers to entry to
mathematicians who want to work on these relevant questions. Specifically, whereas we
discussed the goal of having publicly-available data sets online, we can take that notion
a step further to the goal of having models online which can be run by other researchers
across the globe; in that way, other researchers could use those model outputs as input
into their own, related models. This would bring us dramatically closer to the
development of “models of models” or “systems of systems,” both of which are critical
components of the study of sustainability and the human-environment interactions.
Some effort has already been made in this direction, including at least successful
online collaborative projects to prove mathematical problems such as open theorems
(Castelvecchi, 2010, and Rehmeyer, 2010), as well as the models used for predicting
weather (WRF) but much more can be done.
In particular, we should encourage standard formats for describing models and
expressing their inputs and outputs that would allow for such links. Models must be
documented to be usable by others, but many already are through the academic papers
that their authors publish and present at conferences. What is missing today is an online
version, not of the paper text, but of the functioning models, allowing other researchers
to test and use them.
As with data sets, “live” models can be hosted in various locations, in central
repositories, or on individual researchers’ websites or social networking sites, and can
be open in the sense of making source code available, or, when necessary, can be