This article was originally published by UNHCR Innovation Service, on Medium. Click here to read the original article.
In search of evidence-based, sustainable solutions, JIPS is building a model for safely accessing and analyzing data across organizations including UNHCR.
By Amy Lynn Smith — Independent Writer + Strategist
What do more than 82 million people look like? That’s roughly how many people were forcibly displaced around the world at the end of 2020 — people who have had no choice but to flee their homes due to crises such as persecution, violence in their country, natural disasters, and more. This includes Internally Displaced People (IDPs) who still live in their home country but were forced to relocate. A figure like that is hard to comprehend, but 82 million people could fill one of the world’s largest stadiums over 900 times.
Every one of those numbers represents a single human being. It’s the mandate of the UN Refugee Agency (UNHCR) to ensure their international protection and promote their basic human rights and, in some cases, ease their transition into a new home. But UNHCR must also keep track of migration patterns to know if — and why — people are on the move and what their most pressing needs are in order to provide the most relevant assistance. The best-laid plans for shelter, food, water, and other essentials are made with accurate information about what’s happening on the ground.
The Joint Internal Displacement Profiling Service (JIPS) is an independent, inter-agency project governed by organizations including UNHCR, other UN agencies, and international non-governmental organizations (NGOs) such as the Danish Refugee Council. JIPS is dedicated to supporting long-lasting solutions for IDPs and helping them maintain their dignity in the process. JIPS works with partners including national and local governments, national statistical offices, civil society, and development and humanitarian organizations. With these partners, JIPS uses a profiling process, which is by nature a collaborative endeavor, to create a shared understanding of displacement situations and the way displaced people live. Profiling relies on a variety of approaches to jointly collect and analyze data on displaced populations, their host communities, and others while taking into consideration the economic, political, and social drivers of displacement as well as the places where people live. Ultimately, this information comes together into an evidence base for implementing the most effective possible humanitarian and development responses, advocacy efforts, and national policies that pave the way for these long-lasting solutions.
As valuable as this data is in assisting IDP and refugee populations, it’s vitally important to protect the anonymity of the information, especially in situations of conflict. Profiling involves both technology and the input of people in collaboration, which means there are various actors involved in the data collection and analysis process. Because of the engagement of so many people, responsible data sharing is relevant both among profiling partners and external partners. JIPS, however, was determined to find a more effective way to gather, anonymize, and share the data necessary for those who need relevant information to access it safely and securely.
According to Wilhelmina Welsch, then-Head of Information Management and Innovation and now Acting Coordinator at JIPS, the journey began in 2018, with an initial exploration of a tool called the Statistical Disclosure Control Micro (SDCmicro) that anonymizes data. But it was a long, arduous process, sometimes taking weeks or months to complete the preparation of a single data set — and in humanitarian situations, timeliness matters, even though time is a scarce resource.
There were other issues, too. In many contexts, when JIPS’ support is requested data often needs to be re-collected because partners are reluctant to share what is already available due to privacy concerns. For example, this was the case in Ukraine: The profiling JIPS was supporting needed access to data from a large ministry to proceed, and couldn’t access it because of privacy restrictions.
“I knew there were better solutions out there, but we didn’t have one at that point other than not sharing data,” Welsch says. “This issue became very concrete then, because we had to leave Ukraine without being able to offer a solution.”
Applying innovation to develop smarter strategies
Fortunately, at the same time JIPS was searching for new ideas, UNHCR’s Innovation Service’s Innovation Fund was seeking proposals for projects specific to data and modeling, among other topics. JIPS applied and was granted funding for a project to research and test innovative solutions to incentivize secure, micro-level data sharing among humanitarian organizations.
This project seeks to address the lack of knowledge and trust around methods of data anonymization and sharing, and test whether anonymization alone is sufficient to protect the data privacy of people affected by crises.
According to Rebeca Moreno Jimenez, an Innovation Officer (Data Scientist) with the Innovation Service, partnering with JIPS on this project is of interest for many reasons. She says there is a lot of work to be done in the humanitarian sector on privacy-enhancing technologies to access and analyze sensitive data in as close to real time as possible. Equally important, that data must be protected and there’s a need to build knowledge and trust around data anonymization methods. One central aspect of the project is to apply existing techniques — such as those that are already used in the processing of mobile data — to securely obtain data from sources such as registries of displaced people, censuses, or administrative surveys conducted with IDP populations.
“In humanitarian work, there’s this urgent need to process data very quickly and in the most secure way,” Moreno Jimenez says. “Some of the partners we work with don’t have the technologies or knowledge to do it. So what this project is about is developing a sound methodology and a toolkit — or at least the start of it so that we can keep building — to showcase how it can be done and establish trust.”
Engaging with a team of experts
For the purpose of the project, external partners were brought in for their unique expertise. However, it was important to JIPS and the Innovation Service to work with partners who understand the humanitarian sector, and the unique challenges faced when gathering data.
“An external party doing data collection might not have the same data protection principles we do, so they might expose vulnerable people and do harm,” Moreno Jimenez says. “This work has to be done through a humanitarian protection lens.”
The partners JIPS chose are Alexandre Poltorak, who was brought on as Technical Lead and specializes in software engineering including encryption; Togglecorp, which provides software engineering and data science expertise; and Flowminder, a nonprofit foundation that uses mobile operator, geospatial, and survey data to support humanitarian and development initiatives.
Poltorak says the work began with understanding exactly what JIPS wanted to do. The goal is to gather relevant data without revealing any personal information. Like any innovation project, achieving JIPS’ goals has taken some unexpected turns during the relatively long process of research, prototyping, and testing. Notably, because of constraints involved in using other kinds of data — because organizations including UNHCR were reluctant to share it and the administrative process would have taken too much time — JIPS and their team had to use mobile data, and that’s when Poltorak recognized an important key to moving forward.
“So we could analyze that data in a privacy-preserving way — and provide some of the results to the public and NGOs — I realized it wasn’t a technological problem,” he says. “It was a people problem: the need to communicate with the telecom organizations so they would feel confident we were protecting the privacy of their users.”
The team turned to a solution to make the most of the mobile data they could access.
They used Flowminder’s FlowKit software, which securely allows access to and processing and analysis of de-identified mobile data, to provide a simulation of raw telecom data to mobile network operators (MNOs). FlowKit is an open-source mobile data analytics toolkit that allows MNOs to securely control and monitor access to their aggregated data, but also provides humanitarian and development practitioners with analytical tools developed specifically for their needs, based on Flowminder’s years of experience and scientific research. The team JIPS brought together to tackle this challenge invented the kind of personal information an MNO would have about their users, then documented the data pipeline to prove that only high-level information was extracted, rather than personally identifiable data.
“Using the FlowKit software, we could show the data provider — the telecom — that there were no back doors or data leaking anywhere. The MNOs de-identify the data before it gets into FlowKit, which handles the data processing and aggregation”
— Poltorak explains.
The work done to bring telecom providers onboard so the JIPS team could use their anonymized data was only one step in the process, although an important one.
Having worked in the humanitarian and development space for some time, Flowminder also understands JIPS’ goal of obtaining information from multiple sources, which FlowKit is equipped to do.
“So many organizations only have part of a picture, but they could do so much more if they could link up all of the pieces,” says Jonathan Gray, FlowKit Lead Engineer and Data Analyst at Flowminder. “I think that’s the promise and ambition of this kind of project: to be able to access and combine bits of data from many organizations and do it safely, securely, and effectively without compromising privacy.”
As Gray points out, Flowminder’s mission includes trying to “unlock the value of mobile phone data for humanitarian and development purposes.” But Flowminder recognizes that what JIPS is looking for is even broader in scope.
From Welsch’s perspective, that’s precisely the long-term solution JIPS is seeking. “There’s a lot of contextual data that’s locked up, such as socioeconomic data or data on vulnerabilities,” she says. “This information is crucial because to find durable solutions, we need to understand if people who have been displaced for many years face challenges because of that, or if they face the same challenges as everyone in their communities. For us, the problem we need to solve is to better understand the context. And for that, we need to make the best, most responsible, use of available data.”
To learn more about the journey JIPS and their partners are taking to achieve their goals, watch for part two in this series.