Decision Intelligence in conflict and disaster recovery

In our years building decision intelligence models for domains like banking, telecom, and more, the project that I am most proud of is the work that we did for Liberia in collaboration with The Carter Center.

The basic idea: countries are complex systems. Understanding how to recover after a conflict or disaster can be a particular challenge. Decision makers often end up working accidentally at cross purposes, due to shared, but invisible, mental models of a situation. This often produces unintended negative consequences.

Amongst its sub-Saharan neighbors, Liberia has received particular attention as a model for improvement elsewhere. As in other post-disaster or post-conflict situations, there are nonobvious second- and third-order effects of interventions, and a deep systems understanding of the situation is critical to making the decisions that produce the best outcomes.

Before our project, Ushahidi worked with The Carter Center to build an interactive case map of the country, which is very helpful to understand the legal situation there.  However, it is possible to go beyond data exploration to interactive decision simulation.

I believe that an interactive decision model, especially if it is web-based and informed by real-time data, can be very helpful in “connecting the dots”: understanding how signals can be found in otherwise noisy data, applying “moneyball for government” principles, along with machine learning and predictive analytics, to a particularly wicked problem. This is my vision, and this little project was a start.

This project was conducted before the Ebola crisis. It showed how decision intelligence can shed light on the interactions between Liberia’s legal system, economy, the police force, and more.  In particular, we sought to understand the impact of the Carter Center’s Community Legal Advisor (CLA) project. As a small project, this only produced a preliminary model, from which solid conclusions cannot be drawn.  However, I believe we have contributed a first rung to the ladder required to solve the most difficult problems of our age.

We built our model in the desktop version of World Modeler, produced a video, and wrote a white paper that you can read if you’re looking for a deep dive.  More recently, we built a web-based demonstrator, which you can try out at the bottom of this post.

The key contributions of our work can be viewed as hypotheses, which are that:

  1. A para-professional intervention like The Carter Center’s Community Justice Advisor program represents a highly impactful use of funds (this has been confirmed in a number of randomized controlled trials as well).    See the graph at right for the results of our simulation.
  2. The results of randomized controlled trials can be combined together into a larger systems model to gain a better understanding of, and make better decisions within, a complex situation like Liberia.
  3. There is a “state space shift” effect, where an intervention, if done right, creates a persistent effect to lead to a new a “perpetual motion machine” cycle: specifically improved economic health leads to higher GDP, which leads to greater ability to fund legal, policing, and corruption reduction initiatives, which in turn improves economic health, and so forth. 
  4. Multiple interventions must occur simultaneously to achieve this vicious-to-virtuous-cycle shift.  img_554d4472c0e51In our model, these were improved policing in combination with the Carter Center’s CJA program.  A particular manifestation of this is shown in the graph to the right. It shows that, where CJA investment is delayed, all resources go to policing, and so the economic improvement is delayed compared to the situation illustrated in the graph above.

So below is our web-based model.  It doesn’t quite fit in this blog format, so you’ll either have to scroll around a bit, or might want to click here to see the model on its own page.

To run this simulation:

  1. Read the background and click “Next page” a few times until a page with “Run Model” in the lower right.
  2. Click the “Run Model” button in the lower right (you’ll need to scroll first if you use the one below).
  3. Move the “Total Funding for Justice Aid” slider to create a budget.
  4. Move the percentage sliders in the green button at left.
  5. Press the “Start Simulation” button.
  6. Experiment with different settings of the green sliders.  Note that in particular, with this model, if the “Percentage for Community-based Justice” box is maximized, then this gives the best performance.

Note that this is only a conceptual demonstration of the simulator. The data used in the simulation is fictitious, is not intended to represent reality in any way, and should not be used as the basis for any decisions, nor should any simulation on this page be interpreted as either implicit or explicit claims about any real entities whatsoever.

Patent pending, copyright (c) 2015 Quantellia, LLC.  All rights reserved.

Lorien Pratt

Pratt has been delivering AI and DI solutions for her clients for over 30 years. These include the Human Genome Project, the Colorado Bureau of Investigation, the US Department of Energy, and the Administrative Office of the US Courts. Formerly a computer science professor, Pratt is a popular international speaker and has given two TEDx talks. Her Quantellia team offers AI, DI, and full-stack software solutions to clients worldwide. Previously a leading technology analyst, Pratt has also authored dozens of academic papers, co-edited the book: Learning to Learn, and co-authored the Decision Engineering Primer. Her next book: Link: How Decision Intelligence makes the Invisible Visible (Emerald Press), is in production. With media appearances such as on TechEmergence and GigaOm, Pratt is also listed on the Women Inventors and Innovator’s Mural. Pratt blogs at www.lorienpratt.com.

You may also like...