Explore the conflict forecasts through our interactive dashboard.
Browse our publications to learn more about our forecasts and methodology.
Download data, figures, and source code, or gain access to our API.
RESEARCH & DEVELOPMENT

Current projects

ViEWS collaborates with a number of external actors and renowned research institutes across the world. Below, we present a selection of our ongoing projects.

THE SAHEL PREDICTIVE ANALYTICS PROJECT

in collaboration with UNHCR
About
The Sahel PA project doubles down on the special challenges facing the Sahel across the triple nexus of humanitarian aid, peace-building and development. The project has brought together 23 United Nations entities with partners – including ViEWS – at a number of prominent international research institutes.

EXPANSION TO THE MIDDLE EAST

in collaboration with UN ESCWA
About
In collaboration with the United Nations Economic and Social Council for West Asia (ESCWA), we have developed a new forecasting model that incorporates data of particular importance to the Arab states. Built as an expansion of the standard ViEWS set-up, the model currently covers all of Africa and the Middle East. 

PREDICTING CHANGES IN FATALITIES

in collaboration with the UK FCDO
About
A research project funded by the United Kingdom Foreign Commonwealth and Development Office (UK FCDO). The project sets out to expand the ViEWS system to also forecasting the number of fatalities in armed conflict associated with three different types of political violence (state-based, non-state, and one-sided violence), providing policy-makers and researchers with the ability to quantify the potential impact and intensity of conflicts.

THE COUNTRY EXPERT SURVEY PROJECT

in collaboration with traditional country experts
About
The Country Expert Survey project sets out to complement and contrast the standard ViEWS predictions for Africa with qualitative assessments from over 70 traditional country experts. Based on nearly 400 assessments collected over 2019-2021, the project will release a comprehensive conflict issue dataset in 2022, coupled with a number of articles presenting the dataset, as well as the implicit and explicit predictions that have been extracted from the survey data. 

SOCIETIES AT RISK – predicting the impact of armed conflict on human development

in collaboration with Karolinska Institutet, CRED, ISDC, Barcelona School of Economics, University of Pittsburgh, University of Gothenburg, University of Cambridge 
About
This multi-disciplinary programme funded by Riksbankens Jubileumsfond brings together scholars from economics, epidemiology, political science, and conflict research to study the impacts of armed conflict on human development in much greater detail and comprehensiveness than earlier studies. It takes a risk-analysis perspective, assessing the expected impact as a function of hazard, exposure, and vulnerability, and considers effects at both the macro and micro level on economies, health, water security, political institutions and human rights, and forced migration. Hazard will be modeled through an expansion of ViEWS to also alert observers to particularly detrimental occurrences of violence.
ABOUT VIEWS

ViEWS in numbers

ViEWS is an ambitious early warning system (EWS) at the frontier of research on conflict forecasting. The open-source tool is continuously developed, tested and iteratively improved.
ViEWS covers…
With a country-level performance of…
0
countries
0
prio-grid cells
0
Recall
0
accuracy
0
types of violence
0
levels of analysis
0
precision
Results from the 2019 ensemble, predicting at least one death per month from state-based conflict over 2015-2017 with an ‘optimal threshold’ of 12.6%.
See Hegre et al. (2019).

Advantages of the ViEWS forecasts

PUBLIC AVAILABILITY

All source code and output from the standard ViEWS system is publicly available

ADAPTABILITY

The versatile ViEWS system can be extended and adapted to the needs of each user and organisation. Contact us to learn more about how ViEWS can assist you or your organisation.

state-of-the-art methodology

The ViEWS system is built using state-of-the-art methods and machine learning techniques such as random forests, ensembling, and Extreme Gradient Boosting, which clearly sets it apart from its competitors. 

Performance Transparency

ViEWS offers full transparency about its predictive performance for informed decision-making. See e.g. the out-of-sample evaluations presented in the 2019 and 2021  articles in Journal of Peace Research.

expert knowledge

The ViEWS network is large and interdisciplinary. Beyond the core team, ViEWS benefits from the expert knowledge of research associates at renowned universities and institutions across the globe; from valuable input from users in government ministries, IGOs, and NGOs; and from qualitative assessments from 70+ traditional country experts.

Never miss a data release!

Sign up for the monthly newsletter to receive the latest forecasts and news to your inbox

The forecasting models

The main themes of conflict predictors informing the ViEWS forecasts

Country-level themes

The model ensemble trained to predict conflict at the country level consists of 16 specialised sub-models. They pertain to five key themes of conflict drivers and model features. 

Peace and security

Models informed by numerous measures of conflict and protest history, with data sourced from UCDP, ACLED, and the International Crisis Group. 

Governance

Models capturing the strength of political institutions coupled with comprehensive assessments of levels of democracy. Data is sourced from REIGN and V-Dem. 

CLIMATE

A drought model informed by the REIGN dataset.  

MULTI-FEATURE

A multi-feature model trained on global data. 

Development

Models capturing demographic data from IIASA, as well as development data from the World Development Indicators.  

Conflict history

Conflict history models informed by various measures of conflict history, including e.g. levels of violence, the time and space proximity to the last fatal incidence. 

HUman and natural geography

Models capturing terrain, distance to natural resources, human geography, and local development indicators. 

Multi-feature, cross-level

A multi-feature model trained on global data, and a cross-level model.

Sub-national themes

The model ensemble trained to predict conflict at the sub-national level consists of 11 specialised sub-models that are trained to pick up on local variabilities in conflict risk. They pertain to three key themes of conflict drivers and model features.