WHO IS DR. LINUS BENGTSSON?
When an emergency happens, whether its unexpected like an earth quake, or the result of a steady build up like many of the violent conflicts we have in the world today, people scatter. They scatter from where they live to not only remove themselves from immediate harm, but then to also find and connect with those they care about. This movement presents one of the larger challenges with delivering humanitarian aid to those who need it.
But what if there was a way to detect where people are going, or where they’ve gone – and determine if, or when, they might come back to their homes… without having to physically see them? This is one example of the work of Flowminder, an organization that uses anonymized data from mobile operators and household surveys to help humanitarian agencies and governments understand the movement of people before, during and after an emergency.
Check out this quick video that give a fantastic snapshot of Flowmider’s work after the Nepal earthquake:
On today’s 117th episode of the Terms of Reference Podcast, I speak with Dr. Linus Bengtsson, who is one of the co-founders and the Executive Director of the Flowminder Foundation. Linus is a medical doctor and researcher. In his academic research, Dr. Bengtsson performs evaluations of the use of new digital data types, such as mobile operator data, to measure and monitor development outcomes.
You can connect with Dr. Bengtsson here:
IN TOR 117 YOU’LL LEARN ABOUT:
- Flowminder, a technology based on phone carrier data to map migration and other demographic dynamics.
- Evolving the conversation about privacy. And how data can be used to help humanitarian work without exposing people’s personal information.
- The value of partnerships in developing and deploying global innovation.
- The limits of metadata, or why surveys will exist for a long time.
- The efforts still required in educating governments and officials about statistics.
OUR CONVERSATION INCLUDES THE FOLLOWING:
- GSM Association (GSMA)
- Digicel Haiti
- World Bank
- Gates Foundation
- United Nations World Food Programme (WFP)
- WFP’s VAM and mVAM (mobile Vulnerability Analysis and Mapping)
- 2010 Haiti Earthquake
- Metadata from Telecommunication Services Providers
- 2015 Nepal Earthquake
- 2014 Ebola Outbreak
- Vietnam male HIV epidemic
EPISODE CRIB NOTES
A Stockholm Nonprofit.
Creates partnership with agencies looking to improve program quality.
FF provides data, from surveys or metadata of telecom operators.
FF is a group of academics, performing sets of analyses.
Creates demographic profiles and forecasting. Movement and displacement, particularly after natural disasters.
Three parallel projects
#1. Haiti. Linus was a PhD student working on men with HIV in Vietnam, technology and health issues.
A partner goes to Haiti for relief efforts. Returns after 2 weeks, they talk. People are leaving the city, nobody know where.
Linus suggests tracking their phones for mapping.
They decide to get engineers and contact the Haiti operators. The data is very sensitive, access was not easy.
Linus goes to Haiti. “We can use the information mobile operators collect routinely.”
First time mobile operation data considered for humanitarian purposes.
The need was urgent.
Digicel (Haiti telco) lived the earthquake themselves. They understood the need. But the commercial side was not sold.
FF had a strong academic background to help matters.
Haiti infrastructure collapsed. “Something had to be done.”
FF (fast forward)
FF creates a solution. Data is anonymous.
GSMA released a privacy protection guideline for humanitarian response in the midst of the Ebola outbreak.
FF complies with GSMA guidelines, which involves keeping personal information of the operators’ servers. In general, data that comes out “should not allow you to connect the dots” and identify someone.
FF gets aggregated patterns, which they use for their studies.
They can map geographic movement of phone users. It’s complicated. “People move all the time,” the most vulnerable ones tend to move the most, leaving FF with gaps in their data specifically for the most vulnerable groups. Plus, “they often don’t have phones.”
Before the event, Linus and Co. were still in advocacy mode.
5 years after Haiti, organizations gained understanding of the issues and the “good ways” to approach locational data.
FF’s relationship with Digicel helps their cause.
Haiti relief conferences. “Legal things were cleared out.”
Nepal and FF innovations
Shorten the time to set up the data system in place, including legal issues.
Relationships with providers is critical for better application development.
Getting into tracking diseases takes a different approach where metadata is not enough.
Demographic and health surveys, matched with mobile and census data.
More applications beyond disasters.
Open Data project with dynamic census covering seasonal movement.
Data integration. Working on expanding access, representativity.
New data sources.
Sharpening methods to predict change.
Examples for agencies
FF and World Bank in Haiti, reviewing WB’s status.
Productivity of farming. Haiti’s agriculture is not competitive v. USA’s.
There are mismatches about ideas versus actual location of people.
FF is implementing migration pattern analysis.
The competitive landscape of massive datasets
“I was actually surprised by the lack of camaraderie when I entered the nonprofit sector.”
There are strong and good friendships and alliances.
FF existence predates “Big Data.”
Many players have “too simplistic a view” of data, issues.
A widespread belief that mobile shopping and usage is a robust indicator of consumption patterns, hence living conditions. “This has not been substantiated.”
For small incomes, phone usage cannot be taken as representative.
Key population segments are often ignored (cultural differences?).
Next biggest current disruption
Statistical analysis is becoming a larger part of global health, development.
Not that “we’re getting more serious about science.” Money has not followed.
But there are “more serious” attitudes about statistical evidence. Analysts are part of the community now.
Keeping Linus motivated and funded
Linus’s work doesn’t often involve working with people on the ground.
It takes times for findings to go from being general overviews into decision making tools.
Gates Foundation and World Bank provide funding.
Not core funding, goes from grant to grant, project-specific.
Researchers come following research that interests them, linked to academia.
They “like to do the research but also want to see it used.”
Making Big Data actually useful
When a local, on-the-ground problem is not known or understood, solutions become about the latest “3D graphic tool.” Not that they are not a useful tool to place complex information in a simple to understand way.
“It is all about simplifying the solution.”
The collection, processing and implementation of data are still not standardized. Before that, lots of developing Ministries of health need education, even if the data and methods are simple.
Google keyword search, alerts.
WFP Phone-based surveys.
VAM (Vulnerability Analysis and Mapping Unit) and mVAM (Mobile VAM).
Please share, participate and leave feedback below!
If you have any feedback you’d like to share for me or Linus, please leave your thoughts in the comment section below! I read all of them and will definitely take part in the conversation.
If you have any questions you’d like to ask me directly, head on over to the Ask Stephen section. Don’t be shy! Every question is important and I answer every single one.
And, if you truly enjoyed this episode and want to make sure others know about it, please share it now:
Also, ratings and reviews on iTunes are very helpful. Please take a moment to leave an honest review for The TOR Podcast!