Skip to content
rebecamoreno edited this page Aug 29, 2019 · 3 revisions

Welcome to Project Jetson

This wiki has been created for describing our innovation process for predicting forced displacement with artificial intelligence. Project Jetson was conceived by the Information Management (IM) team in UNHCR Somalia operation and developed by UNHCR Innovation Service starting mid-year 2017.

Forced displacement, is a complex phenomenon. Attempting to predict population flow of those who are internally displaced (IDPs) and/or crossing an international border (refugees) to save their lives is a challenging task that can be undertaken using a variety of tools. We, as Innovation team, decided to undertake this journey by exploring the capabilities of data science and artificial intelligence to attempt predicting the complexity of forced displacement and with that, the agency of individuals behind their decision to flee.

The "recipes" we built in this wiki are driven by Somalia operation: the idea of predicting forced displacement in advance in order to strengthen emergency preparedness and operational response. The first one is attempting to predict arrivals with one month in advance (recipe/experiment #1) and the second one trying to do it with three months in advance (recipe/experiment #2). Finally experiment #2 also portrays a solution for automating one month in advance predictions.

Acknowledgement

To our UNHCR colleagues and humanitarian partners: Thanks for sharing your operational context knowledge of the region with us. Special recognition is to given to UNHCR Somalia team and to the colleagues working with the Protection and Return Monitoring Network (PRMN) collecting data in the most remote places in Somalia. Additionally we would like to give special recognition to the UNHCR Somalia operational data management team based in Nairobi and to the UNHCR Dollo Ado, Melkadida-sub office for sharing their data every month. Last but not least, to the IM UNHCR East Africa Regional hub team for providing feedback on the project.

To our development agencies partners: special thanks to WMO and ICPAC team and to the FAO SWALIM and FAO FSNAU teams, as well as FAO Innovation team. Their data and technical knowledge on climate, weather and market prices is key for the development of this project. This data represents those influential factors for movement in an operational context where those who are forcibly displaced, highly depend on them.

To those peacebuilding institutions: working on documenting and publicly sharing their data and knowledge about violent conflict. Special thanks to ACLED Data colleagues for their full understanding on the importance of open data for humanitarian response. They probably are not even aware we are using their datasets, and how valuable they are to our operational and research work.

To our data science and artificial intelligence mentors: UN Global Pulse colleagues and their Data Fellows Program Thanks for your patience and guidance in this project and all your inputs to develop it. Also thanks to Uptake Foundation for their data fellows program and their technical training provided to our team. Thanks also to OCHA Humanitarian Data Centre for the opportunity to portray Jetson as one of the first experiments in the humanitarian community for predictive analytics.

To the academic community: The University of Essex Human Rights, Big Data & Technology Project (HRBDT) has expanded our horizon research to create a spin-off to the project: the inclusion of satellite imagery for understanding the interrelation between weather/climate anomalies, conflict and displacement. Thanks for providing your knowledge and human resources for that. This collaboration has recently being expanded by joining forces with Omdena - a global group of volunteer AI engineers that work to solve specific challenges with AI - by developing a Displacement Challenge for analyzing satellite imagery.

But most importantly: special thanks to those refugees who provided us with feedback on our project in Dollo Ado. They might not know their struggle to flee and their personal stories helped us validated our scientific assumptions. Special recognition to all those more than 70 million people who are forcibly displaced around the world. Our work is for you.

Clone this wiki locally