Early Warning & Forecasting Hub

Explore the Early Warning Tools

Country Change Level Map

The Country Change Level map identifies countries and territories experiencing a change in political violence and demonstrations in the current week relative to a historical norm. This map is displayed if the “Country” view is selected from the “Map & Table View” menu. If “Previous month” is selected in the menu, the historical norm is the most recent four weeks prior to the current week; if “Previous year” is selected, the historical norm is the most recent 52 weeks prior to the current week. 

Change levels are calculated using the number of events given the selected event and/or sub-event types in the sidebar menu. The colors and their corresponding categories and thresholds are as follows:

Color	Category	Percentage Change Threshold
	Large increase in activity	Greater than a 25% increase
	Moderate increase in activity	Between a 10% and 25% increase
	Limited change in activity	Between a -25% decrease and a 10% increase
	Large decrease in activity	Less than a -25% decrease

Countries with two events or fewer on average in the previous month/year and five events or fewer in the most recent week of ACLED data are categorized as “Limited change in activity” to account for low baseline activity.

Please note that these maps depict change in violence and demonstration rates, not overall levels. For example, a very violent country that experiences a significant decline in violence may still be very violent, but it will register a “large decrease in violence” on the map.

Event Locations Map

The Event Locations map shows events of the selected event and/or sub-event types for the latest week of ACLED data. This map is displayed if the “Subnational” view is selected from the “Map & Table View” menu. If there are many events near a location, the points will cluster into a single circle with a label denoting the number of events near that location. Clicking a cluster will reveal individual events near that location or additional sub-clusters within the larger cluster. Hovering over the individual points display event information, like the location, date, actors, and event/sub-event types.

Event locations styled with a ◆ indicate that Emerging Actors have been identified. These are actors that were active in the latest week but not within the last year. For easy identification, Emerging Actor events are always shown – they do not cluster with other nearby events on the map like those without Emerging Actors. When hovering over an Emerging Actor event, a ◆ flag will appear next to the actor(s) that are emergent in the event.

The individual points are colored by event type, with the included types selected in the sidebar menu.

Event Count and Change Statistics Table

The Event Count and Change Statistics table shows the latest week’s event count for the selected event and/or sub-event types. If the “Country” view is selected from the sidebar menu, the table summarizes activity for each country and territory in the world. If the “Subnational” view is selected, each row in the table is for a single ADMIN1 within a country or territory. If a specific country is selected from the sidebar menu, this table retains only ADMIN1s within that country.

The table displays three statistics for each entry:

Column Name	Description
Weekly Count	Total number of events for each country or ADMIN1 for the latest week of ACLED data.
Four-Week Average/Yearly Average	The average number of events for a corresponding country or ADMIN1 over the most recent four weeks (if “Previous month” is selected) or 52 weeks (if “Previous year” is selected).
Pct Change	The relative difference between the most recent week’s event total compared to the selected average, expressed in percentage terms. The colors correspond to the change levels categories.

The table is interactive, with complete sorting and filtering capabilities. Click a column header to sort by that column. To search for a specific country or ADMIN1, type into the bars under those headers. The bars under the statistic column headers may be used to filter values, retaining only rows with greater than/equal to the specified value. Clicking a specific country or ADMIN1 will select that unit in the sidebar menu, updating the other elements of the dashboard to reference only that selection. To export the contents of the table to a spreadsheet, click the “Download” button.

The colors for the “Pct Change” column correspond to the following thresholds:

Color	Category	Percentage Change Threshold
	Large increase in activity	Greater than a 25% increase
	Moderate increase in activity	Between a 10% and 25% increase
	Limited change in activity	Between a -25% decrease and a 10% increase
	Large decrease in activity	Less than a -25% decrease

To adjust for low baseline activity levels when calculating percentage changes, countries with two events or fewer on average and five events or fewer in the most recent week of ACLED data are categorized as “Limited change in activity.” ADMIN1 units with one event or fewer on average and three events or fewer in the most recent week of ACLED data are similarly categorized as “Limited change in activity.”

Applies to…	Average events in previous month/year	Event count from most recent week of ACLED data	Category
Countries	2 events or fewer	5 events or fewer	Limited change in activity
ADMIN1s	1 event or fewer	2-3 events	Limited change in activity

Actor Network

The Actor Network graph visualizes interactions among the 25 most active actors for the selected event and/or sub-event types for the most recent week of ACLED data. Each node is either a single actor or a group of actors of the same actor type. A connection between icons implies that the actors interacted in the same event(s). 

The contents of the graph depend on the selections in the sidebar menu:

• If “All” countries and territories are selected, the network shows interactions for the entire world. 
• If a single country or territory is selected while “All” ADMIN1s are selected within that country/territory, the network shows interactions for that entire country.
• If a specific ADMIN1 within a country/territory is selected, only interactions within that ADMIN1 are shown in the graph. 

Identifying actors

• Actors are colored by their type and sized by the number of events in which they were involved in.
• Emerging Actors – denoted by a ◆ – are actors that were involved in an event in the most recent week of data but not active during any other week in the past year.
• Hover over a node to reveal an actor’s name and associated event count.
• Select a specific actor from the network menu for quick identification within the network.

Actor groupings:

• Each Emerging Actor is plotted as a single node, regardless of how many events they were involved in.
• For the Global Actor Network, non-Emerging Actors are grouped together by actor type.
• When an individual country or ADMIN1 is selected, the 25 most active actors, in addition to all Emerging Actors, are plotted as single nodes. All other actors are grouped together by actor type.

Why focus on emergent actors? 

The presence of new actors, and those that move from latent to active operations, points to a change in the conflict environment. In areas that are relatively peaceful, it indicates that the incentives and opportunities for violence have shifted. In high activity spaces, new or re-emergent conflict agents can indicate factionalization of existing opposition groups; new attempts by governments to seed the formation of militias; and new interest groups engaging in violence with an agenda to upset the status quo, among other trends. When areas become incubators for new groups, there is a high possibility that more groups may emerge. It is probable that the frequency and intensity of actions will increase, or at best, that more groups will engage in activity. This means that engaging in this environment, or seeking peace, will be complicated by ever more agendas and compromises with armed groups.

Civilians often bear the burden of violent activity involving new and re-emergent groups. Such actors are often weaker than those that are already established, and in order to demonstrate their role without engaging in combat, these groups may focus on civilians as they are easier targets and because attacks on communities can generate significant media coverage and public attention, elevating the presence of that group in a multi-agent environment.

Trends Chart

The Trends chart shows events per week over the last year, situating current activity within the historical context. When “All” countries and territories are selected, the chart shows the total number of events per week across the world. If an entire country is selected, the event total is for the entire country. If a specific ADMIN1 is selected, the event total is just for that ADMIN1.

Each bar represents the event count for a given week over the past year. Weeks are shown based on ACLED coding weeks, which begin on Saturday and end on the following Friday. The trendline shows the moving average event count, looking back to the last four weeks (if “Previous month” is selected) or the last 52 weeks (if “Previous year” is selected). The four-week average is useful for understanding how current activity relates to the short-term norm, while the yearly moving average is better suited for comparing to a longer-term norm.

Weekly event counts are flagged as “High Activity Anomalies” if they are two or more standard deviations above the moving average. The total number of high activity weeks during the past year is annotated below the chart.