MSc Thesis graduate student Food security & Malnutrition Machine Learning Model

Position title: MSc Thesis graduate student Improving an open data and machine learning model to predict the food insecurity and malnutrition status of vulnerable communities.
Duty station: The Hague and Utrecht (50-50)
Time period: 4-6 months fulltime (depending on your institution’s requirements)
Status: Student


Since 2017, the Netherlands Red Cross 510 data initiative and the ICCO Cooperation cooperate to make better use of data in coping with disasters and to improve our development programs in developing countries. Together, we act before, during and after disasters to meet the needs and improve the lives of vulnerable people. We do this irrespective of nationality, race, religious beliefs, class or political opinion. The cooperation between 510 and ICCO Cooperation focuses on food security and nutrition and is supported by the Dutch Coalition for Humanitarian Innovation (DCHI) and the Sight and Life Foundation.

In 2017, one of our MSc students wrote a thesis about “Indicating Food Security in Ethiopia with Open Data and Machine Learning”. This student made a first model to predict the food security situation at the lowest geographical level possible by combining various open data sets. Food security predictions can add to the Community Risk Assessment model of the Netherlands Red Cross. Overall research questions were: How can information regarding food security and nutrition be further integrated into a holistic understanding of vulnerability of communities supporting humanitarian and development decision makers before, during, and after interventions? How can open secondary data that is available throughout the country be correlated to and supplemented by detailed assessment data regarding food security and nutrition that are available from smaller geographical areas in a few places? A short explanation and the Thesis can be found through:

Purpose of the position

To validate and improve the accuracy and robustness of the model to predict food insecurity. This technical position requires a student with a strong statistical learning and methodological background.

Research questions and activities
Future research regarding data input and output requiring the student to find, clean and collate additional open data as well as to work with ICCO and 510 to open existing detailed assessment data:
• Output data:
o Try to find out in more detail how FEWSnet Current Situation Assessment is determined.
o How can HFIAS data from ICCO (or more in general on-the-ground survey data at output level) be used?
• Input data:
o Analyze if additional (open) data sets can be added for the input of the model; same independent variables but data on these variables with higher quality/granularity.
o Include other independent variables such as on Climate and environmental stressors: additional data from e.g. and Social and political unrest or war.

Future research regarding machine learning model, such as:
• Assess sample bias and variable importance, use of dummy variables.
• Use more test- trainset combinations ( k fold cross validation).

Extending the model:
• Test and adopt the model for other countries (most likely Kenya and Uganda).
• Test if (open) (mal)nutrition data can be added to the model.
Assess how the model can be integrated in 510’s Community Risk Assessment dashboard (mostly from a technical point of view, as another MSc student works on the policy management aspects).

• Bringing your knowledge, experience and passions.
• Affinity with – and motivated by – humanitarian and development work.
• Currently following a University Master’s degree, has excellent grades at the university and is in the phase of doing an internship or doing a masters graduation project.
• Preferably in the field of Computer Science, Econometrics or otherwise strong statistical background.
• Is fluent in English.

We offer
A fulltime graduation position, with 50% in the 510 office in the Hague and 50% with ICCO in Utrecht. We offer travel compensation to our offices. You will be part of a vibrant community of data-driven humanitarians. We share and teach one another where possible.

The post will be open until the right candidate has been found. Please reply to this advert with:
1. Why you’re interested in doing this graduation project in less than 100 words.
2. What currently interests you in less than 100 words.
3. A TED talk that you thought was particularly interesting.
4. Your CV.

Please send these 4 items to Marc van den Homberg, and Martijn Marijnis,

About 510 and its place within the Netherlands Red Cross
510 is a self-organizing data innovation initiative of the Netherlands Red Cross. Our vision is that smart use of (big) data will help towards faster and more (cost) effective humanitarian aid at a global level. Contributing to open data, data analysis and capacity building in governments and NGOs are essential to increase the understanding of humanitarian data. We want to shape the future of humanitarian aid by converting data into understanding, and to put it in the hands of humanitarian relief workers, decision makers and people affected, so that they can better prepare for and cope with disasters and crises.

The 510 initiative was established early 2016 and has grown since into a team of over 40 core and project staff, graduates, students, trainees and volunteer data experts with diverse backgrounds. The team is developing data and digital solutions for both international and domestic humanitarian aid. Our main results are published on

510 works together with all departments of the Netherlands Red Cross, as well as with different teams in the International Federation of Red Cross and Red Crescent Societies (IFRC), other Red Cross National Societies, Universities in The Netherlands and abroad, United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA), national governments and through the Dutch Coalition for Humanitarian Innovation (DCHI) with other NGOs and businesses.

We work in virtual teams using a fully digital collaboration environment, which enables us to bridge time differences and physical distances and to continuously work together on some of the most difficult humanitarian challenges. However, we also greatly value the opportunity to work together in a face-to-face setting, which is possible in our office in The Hague.

About ICCO Cooperation
ICCO Cooperation is the interchurch organization for development cooperation with a strong focus on economic development, food security and livelihoods and always working in multi stakeholder partnerships. Through the ACT-alliance membership, ICCO Cooperation is able to gain access to numerous UN bodies, effective advocacy channels and campaigns, joint fund raising and coordinated humanitarian relief (incl. key users of exposure data sets with socio-economic indicators relevant for this challenge).

ICCO Cooperation is gathering household level data in various countries using various indicators including the HFIAS (Household Food Insecurity Access Scale) and DDS (dietary diversity score). In many cases we are using the Akvo FLOW application to do so. Data is gathered for problem analysis and for monitoring, evaluation and learning purposes. Among other social domains, ICCO Cooperation has a particular interest in food security. ICCO Cooperation reckons that gathering primary data is costly and time intensive and hence would like to consider alternative solutions or add-ons.

Comments are closed