Data Into Action

Charles DeTar, a C4 researcher, has put together a fantastic navigator for the Afghan War Diaries data published by Wikileaks.  If you have not had the time or energy to look through the diaries, this is a great way to piece together events from multiple entries, to search and browse based on type of event, etc.  Incredible that he threw it together in just a few days.

If you like it, vote it up!

mefi: http://ur1.ca/0zeqa , reddit: http://ur1.ca/0zeqc

Data into action:  Action!

How does one transform data into social action?  What combinations of information, access, interface, interaction, and social organization best lead to substantial social transformation?  These questions are at the heart of our research at the MIT Center for Future Civic Media, and I had the honor of posing them to three top thinkers and practitioners in data and social change during a plenary panel at our conference last month.

In challenging the speakers, I gave two models as examples.  I will call the first model the “Brockovitch”.  In it, a[n] [investigative journalist / whistleblower / activist / regulator / informed citizen] carefully researches and analyzes an obscure database that was probably generated by the government to comply with a reporting law passed years before.  The investigator then uses [statistics / data visualization] to [sue / advocate / write a story about] [systemic discrimination / corruption / inefficiency / a sex scandal].  This is a well known model, off-gassing heroism: the truth is out there, follow the money, look for patterns, break it wide open.  Slap the analysis into the [evening edition / indictment / policy brief / public domain], and use it the proof to elicit change.

The second model is more quotidian, widely distributed, and decidedly less cinematically heroic; I will call the “Tufte”.  In the second model, data that has long been available [realtime bus GPS data / credit reports / local EPA toxics inventories] but difficult to access are suddenly made legible through [information design / just in time delivery / phone ‘app’ ].  Unlike the first model, there was no conspiracy to uncover; rather, useful information was simply not making its way to the citizen in a way that they could use when they needed it.

I would like to add a third model, one that seems to be showing a lot of promise, which I will call the “LittleSis”.  In this model, no suitable database exists for a particular application.  But through [aggregation / mashing up] some basic data or categories, [informed citizens / specialists / retirees / activists] collaboratively research and collect data, using a [wiki / web application / edge-node database].  At that point the data will be used similarly to the “Brokovitch,” perhaps mined by journalists or regulators.  Unlike the “Brokovitch,” however, users played just as important a role in generating and contributing data as in analyzing and leveraging it.

It should be clear from these three models that there are many parameters involved.  Who are the users?  Is the data governmental and top-down, or is it being generated by users?  Once the data is analyzed, what is the next step?  Does the same entity host both the data and the system that makes the data useful?  Indeed, there are so many parameters that it could easily be argued that these three models represent entirely different activities, and there is no similarity between them.

Certainly there are nuances in each of these models, nuances that necessitate different sorts of design and implementation.  But I would argue that they have many features in common, and indeed represent points on a spectrum.  For example, if a strong site like LittleSis were to do such a good job of identifying conspiratorial connections and oligarchic subterfuge that lobbyists had a harder time creating corruption in government, there would be less need for journalists to have to muckrake through the databases at a later date.  And if real-time bus data was available as a convenience for riders waiting in the rain, then data would automatically available to research inequity of transportation routes based on neighborhood income.  Each of these models represents an appropriate fulcrum point at which well-marshaled data can be leveraged to assist civil society.  But which model is likely to be the most effective in a particular situation?  If we integrate these models vertically, will that lead to more or less impact?  Are there 50 more models out there waiting to be discovered?

While each expert on the Data Into Action panel answered the question based on their extensive experience, all the panelists — Ellen Miller (Sunlight Foundation), Nick Grossman (Civic Works & OpenPlans), and Laurel Ruma (O’Reilly and Associates) — indicated that there is still a lot of experimentation to be done before anyone can answer the question with any certainty.

Then by all means, let us experiment!   This list will mix social entrepreneurs, academics, journalists, government functionaries, and activists.  Let us use this blog and this mailing list to share our experiments and experiences, to help and challenge each other, to build our collective knowledge, and to drive our ability to push information technologies to build a more open, equal, and informed society.