VIEWS3 methodology

Overview of the methodology and data infrastructure behind the continuous forecasting models in use from 2022 onwards
The current iteration of the VIEWS system is composed of an open-source data infrastructure known as VIEWS3 with a web-based CLI called viewser, under which the  fatalities conflict prediction models are trained and calibrated.

OVERVIEW OF THE PREDICTION MODEL
The fatalities model

VIEWS is an ambitious early-warning system at the frontier of research on conflict forecasting. The open-source tool is continuously developed, tested and iteratively improved, resulting in frequent releases of new models and versions thereof.
The current model is known as the fatalities model. For each month in a  3-year forecasting window, it generates predictions for the number of fatalities in impending conflict, as well as dichotomous forecasts for the probability of at least 25 battle-related deaths (BRDs) per country-month and at least 1 BRD per PRIO-GRID-month. 
Changes to the fatalities model over time are documented in our version-tagged source code on GitHub, where the complete model documentation is accompanied by a written model changelog. 
Current model: the fatalities model
Model version: 001 (002 coming soon)
  • Hover over the boxes for more information.

UPDATE SCHEDULE

Monthly

PREDICTED TYPE(S) OF VIOLENCE

State-based conflict

FORECASTING WINDOW

1-36 months ahead

PREDICTED OUTCOMES

Continuous & dichotomous predictions

COUNTRY-LEVEL COVERAGE

Global

SUB-NATIONAL COVERAGE

Africa and the Middle East

Read the technical report presenting the first version of the fatalities model (fatalities001)

Predicting fatalities

Håvard Hegre, Forogh Akbari, Mihai Croicu, James Dale, Tim Gåsste, Remco Jansen, Peder Landsverk, Maxine Leis, Angelica Lindqvist-McGowan, Hannes Mueller, Malika Rakhmankulova, David Randahl, Christopher Rauh, Espen Geelmuyden Rød & Paola Vesco Report, Uppsala University, 9 June 2022

ABOUT THE INPUT DATA
The features informing the VIEWS models

The fatalities model is informed by data on hundreds of variables from data providers such as the Uppsala Conflict Data Program (UCDP), ACLED, PRIO-GRID, the World Bank, IMF, FAO, Mapsdam, SPEI and MIRCA. Based on these raw data variables, VIEWS also creates a suite of additional variables by applying data transformations such as time and space lags, imputations to fill in for missing data, and other common data processing techniques. Together, the raw- and processed data variables informing the VIEWS models are referred to as  features, which are grouped into feature sets based on the overall theme they relate to and/or the data provider(s) from which they are derived.
The feature sets informing the sub-models and model ensembles in the fatalities  model are documented in the GitHub repository viewsforecasting, coupled with notations of any data transformations that have been applied to the constituent features. The data transformations, in turn, are described in a dedicated Jupyter notebook. 
Why categorize data into feature sets?
Categorizing input data variables into feature sets is part of the standard data organization routines in VIEWS, which greatly facilitates model development. Amongst other benefits, it allows us to call upon a pre-determined set of features, which is maintained in a single location, when training our models. This minimizes the risk of human error when compiling the input datasets and greatly facilitates maintenance of the model documentation. 
  • Did you know?
Data relating to each feature set can easily be queried from the VIEWS database by means of the web-based CLI viewser that interacts with the backend data infrastructure VIEWS3. This is, however, currently limited to authorized users.
If you or your team seek access to our input and replication data, please contact the VIEWS team for personal assistance. 

Key feature sets

Conflict history

A suite of features capturing the history of conflict in each country and sub-national grid cell, e.g. the number of battle-related deaths per unit and level of analysis, and measures of the temporal and spatial distance to recent conflict events. 
Data providers: Uppsala Conflict Data Program (UCDP), The Armed Conflict Location & Event Data Project (ACLED).

Political institutions, democracy

Features that capture democracy indices and the strength of political institutions in each country, such as liberal democracy, rule of law, equality, and the level of exclusion of social groups in politics. 
Data providers: Varieties of Democracy (V-Dem)
Measures of development as provided by the World Bank Indicators, e.g. GDP per capita, infant mortality rate, and school enrollment. 
Data providers: The World Bank (Word Development Indicators, WDI)

Economic growth

A feature set focusing specifically on historic and future economic growth, e.g. real GDP growth per year and growth forecasts for the coming years. 
Data providers: The International Monetary Fund World Economic Outlook (IMF WEO)

Climate & societal vulnerability

Feature sets capturing climate extremes and societal vulnerability to climate hazards and other external shocks, e.g. climate extreme indices, reliance on agriculture, crop yields, precipitation, freshwater withdrawal, water management efficiency, and access to renewable resources. 
Data providers: United Nations Food and Agriculture Organisation (FAO), FAO AQUASTAT, PRIO-GRID, MIRCA, MAPSPAM, SPEI Global Drought Monitor

News monitoring

A feature set based on the Mueller & Rauh (2018) topic model, which captures conflict risks as drawn from a topic analysis of news media. 
Data providers:  Mueller & Rauh (2018)

Natural and social geography

A feature set capturing terrain type, distance to natural resources, demography, proximity to cities and country borders.
Data providers: PRIO-GRID

Food security and access to basic needs

Feature sets capturing staple food prices along with measures of food security and access to basic human needs, such as mean food prices, food price inflation, undernourishment, access to clean water, and basic sanitation. 
Data providers: United Nations Food and Agriculture Organisation (FAO), FAOSTAT

GENERATING THE FORECASTS
Model training procedures

Step
1

Sub-models: combinations of feature sets and algorithms

As a first step when training the VIEWS models, each feature set is paired with an advanced machine learning algorithm.
The fatalities  model currently employs four such algorithms, which you can read more about in our technical reports on the model: random forests, gradient boosting, markov models, and hurdle models. 
The result is a series of sub-models, or constituent models are they are more commonly called, that use patterns in their respective subsets of historic data to generate predictions for future conflict.
Step
2

Ensemble models: groups of sub-models using “the wisdom of the crowd”

Much like a crowd tends to be wiser than the individuals composing it, prediction models that are informed by a number of smaller and specialized sub-models are known to be more robust and generate stronger predictions than single models.
As a second step in the model training procedures, the sub-models above are therefore combined into two groups or ensembles of models – one ensemble for each level of analysis. 
Two different ensembling techniques are used for this purpose:
  • The country-level ensemble model combines the predictions from each of the sub-models using a genetic algorithm that assigns different weights to the contribution from each model in order to maximise predictive performance.

  • The sub-national ensemble model, in turn, uses a simple unweighted average of the sub-model results.
The ensembling techniques above are motivated and described at length in the technical report on the  fatalities model. 

ABOUT VIEWS3 & VIEWSER
The data infrastructure supporting the fatalities models

The fatalities models are built in a brand new sophisticated data infrastructure called VIEWS3 – the third iteration of the back-end system and database supporting the VIEWS prediction models. Users can interact with the VIEWS3 system using a web-based CLI called viewser, albeit such use requires a database certificate/special access.  VIEWS3 and viewser are documented in a suite of open-source GitHub repositories that users are welcome to consult for detailed information.