Visual style guide

This visual style guide aims to ensure consistency across all ACLED Analysis visual outputs. This guide is directed to staff from the Analysis, Data Science/Management, External Affairs, and Grants & Development departments involved in the production of data visualization.

The guide consists of two sections. The first section outlines the brand identity policy. The second section introduces guidelines for planning and selecting visualizations and best practices to keep in mind when building graphic assets.

When producing visual outputs, all authors should seek to apply the following rules:

Consistency is essential

A consistent visual identity gives ACLED outputs a unique stamp that helps users to identify them anywhere.

Less is more

Use a minimalistic approach in data visualization. Only include what is truly needed.

Put yourself in the reader’s position

Ask yourself: is too much time needed to read and understand the visual?

Focus on the visual elements rather than the text

Avoid adding too large titles, footnotes, or annotations. Be as brief as possible when it comes to text.

Visual Identity

Data Visualization Guidance

Planning visualizations

Choosing the correct visualization

Improving visualizations

Visual Identity

Visual_style_guide_(February_2026).pdf

Data Visualization Guidance

This section introduces you to the best practices for producing accurate and compelling visualizations, from a single chart for a report to a detailed infographic. The contents are structured for you to think about data visualization as a stepped process: planning, choosing, and executing.

Planning visualizations

An effective visual is the result of a planning process. Planning not only helps deliver a better result but also makes the process of achieving it more efficient. This section introduces a series of tips to keep in mind when planning visuals for written reports or infographics.

Planning visuals on reports

Before filling out the Graphics Pack for your report, ask yourself these four questions to ensure that the visual you are requesting is relevant:

Does this section of the text need a supporting visual?
What type of visual is the best option?
If I don’t include a visual, does the argument still remain clear enough?
Does the visual need to be interactive or static? Does having an interactive visual make the argument clearer compared to having a static one?

Planning infographics

In this section, you will find some tips to help you during the drafting process.

Draft a mockup of the infographic

The mockup is a draw of the infographic without any data or actual text. It is helpful to identify which charts, maps, text boxes, and overall visual elements should be included and where and how they should be displayed.

Why are mockups so important?

Mockups help save time. The creative process only happens at the beginning, so you don’t have to improvise or continuously think of new ways to present information.

Mockups make the review process more straightforward. As the general contents of the infographic are known beforehand by the reviewer, the piece won’t need to get through major changes unless something extraordinary happens.

Keep text as concise as possible

An infographic relies on visual elements rather than text. Text should only be used to present key ideas and provide context for the visuals. A text box should have no more than two or three sentences, and you should have no more than three text boxes in an infographic.

What visual elements can I use?

Key figures, icons, charts, and maps should be used in an infographic. Please notice that you don’t have to use them all simultaneously, only those that are relevant to support your thesis. In some cases, using them all creates clutter.

Choosing the correct visualization

Chart types

If you are unsure as to which visual you should use to support your argument, you can use this table to find potential options. It is arranged to display the different types of charts that can be incorporated according to the type of trend or pattern that you are exploring.

What do you want to show?

Change across time

Example: April and May 2022 saw the lowest levels of total reported fatalities since the start of ACLED’s Yemen coverage in January 2015. Throughout these two months, the weekly total fatality counts were consistently the lowest since March 2015.

The UN-Mediated Truce in Yemen.

Impacts of the First Two Months

Line chart

Bar chart

Comparison between categories

Example: Comparing Boogaloo actors with the Proud Boys and the Oath Keepers, the scale of violence comes into focus. This visual shows the major disparity between the number of events involving Boogaloo actors and the reported fatalities associated with these events.

Actor Profile: Boogaloo Boys

Bar chart

Part-to-whole

Example: The Wagner Group engages in high levels of civilian targeting in both CAR and Mali. Civilian targeting accounts for 52% and 71% of Wagner involvement in political violence in CAR and Mali, respectively.

Wagner Group Operations in Africa.

Civilian Targeting Trends in the Central African Republic and Mali

Stacked bar chart

Donut chart

Pie chart

Correlation

Example: A combined look at analysis of ACLED data and Media Matters analysis of CRT mentions by Fox News shows a powerful correlation between increased demonstration activity and increased negative media coverage.

Fact Sheet: Demonstrations over Critical Race Theory in the United States

Combined column/line chart

Map types

According to the granularity and the trend you are analyzing, you can choose between a point map (using latitude and longitude) or a choropleth (coloring countries or administrative divisions).

This table can help you identify the most suitable map for your purposes. Please consider that the table only works when you want a simple map to show a geographic trend using ACLED data only. Later in the section, you will find tips for advanced mapping.

How to use the table?

Start by identifying the scope of your analysis: Multiple countries? Multiple Admin1s within a country? Specific areas within an Admin1, like Admin2, Admin3, or Locations? Select the correct option from the first column.
From the variable you are using (analysis categories, event/sub-event types, actors, etc.), identify whether you want single or multiple values. Select the correct option from the first row.
The match between the selected values will show the most appropriate map for your analysis.

Examples:

A map depicting the number of political violence events in the Middle East during November 2022.

Combination: Worldwide/Regional (Middle East - a region) + Single (Political Violence - a single analysis category)

Map type: choropleth

A map displaying the number of violence targeting civilians (VTC) events in the Zulia state (Venezuela) during January 2019, disaggregated by sub-event types.

Combination: Subnational (Zulia state - an Admin1) + Multiple (VTC sub-event types)

Map type: point map

Mapping guidance

For regional and countrywide trends, choropleth maps take precedence over point maps. Point maps should only be used in these cases if the argument supported by the visual requires data disaggregation by event, sub-event types, actors, etc., or to display areas of activity/operation.
In point maps, point sizing by the number of events or reported fatalities should only be used when the argument explicitly references intensity or activity levels. Point sizing is not needed if it is used to display areas of activity/operation.
If there are clear geographic trends within a region/country or areas with limited or no activity at all, use a zoom-in map to help focus the reader’s attention. It also helps to reduce the clutter associated with showing areas with no activity or relevant trends.
In a zoom-in/subnational map, include a second zoom-out map on one of the corners, displaying that area within the country/larger region (see an example in the image below.)

Do not include Antarctica and the Arctic in global maps, as those regions are not relevant for visualizing conflict data.

Advanced Mapping

Combing ACLED data with external datasets

In some cases, information from other datasets can provide context to trends in ACLED data or help establish correlations between ACLED data and other phenomena. Combining multiple datasets often requires combinations of map types. Here are some suggestions that can work for different scenarios.

Combine point and choropleth maps: overlap a choropleth map with the external data and a point map with ACLED data. This might be the most common choice since many datasets don’t offer geographical information at a high level of granularity.

For example, in ACLED’s report about political violence during Brazil’s 2022 presidential runoff, ACLED data were combined with election results data.

Combine two point maps: this option works when the external data has information on key locations or selected points. Use this option only when the size of the external data is smaller than ACLED’s or vice versa in order to make the visualization easier to understand.

You can use this type of combination when you want to observe political violence activity around critical infrastructure for which you have exact coordinates, such as military bases, airports, government buildings, healthcare facilities, etc.

If you are unsure about which one to use or if none of the types work for your analysis, contact the Research Analyst assigned to support you.

Interactive maps

Interactive maps are good options when the reader needs various types of information to better understand an argument around a geographical trend. Before requesting one, please ask yourself the following questions:

Does my argument need a map showing geographical patterns and other trends?
What should be the conclusion that the reader should reach after interacting with the map?
Will the map be easier to understand if the user can interact with it?
If using interactions, what type should the map have? Buttons to turn on/off features? Sliders? Tabs? Tooltips?

Data disaggregation

Charts and maps can include data disaggregated by event/sub-event types, actors, etc., only if your argument refers to the effect/behavior of a third variable in/regarding the trend you are describing.

In the table below, you will find two examples of scenarios when data disaggregation is needed and when it is not.

Violence targeting civilians has increased during the last three years.

Notice that the argument is referencing violence targeting civilians only. Therefore, it is not necessary to disaggregate by sub-event types to illustrate the events increase.

Violence targeting civilians has increased during the last three years. Most of this violence has taken the form of attacks.

Notice that the argument mentions an increase in civilian targeting and the most prominent sub-event type. In this case, we need to disaggregate the bars by sub-event type to display both trends.

Improving visualizations

By default, most software displays plots with many added (and often unnecessary) elements, known informally as clutter. Much of these elements can be removed or edited to get a more comprehensible and visually attractive outcome.

When simplifying elements of a chart, always consider these principles:

Remove elements that do not transmit information (i.e., unnecessary colors, gridlines, 3D effects). In other words, the elements that remain unaltered even when data changes.
Remove elements that, although they change when data also changes, can be replaced by simpler options. This is the case of an axis, which often can be completely erased or replaced by data labels. However, the editing of these elements has to be made with particular caution, as their removal may not be appropriate in all situations. For example, removing the y-axis in a line plot and replacing it with labels doesn't work for large time periods.
Leave enough blank space between elements to allow readers to take “breaks,” but avoid leaving too large empty areas.

Consider this example of a bar plot representing data from political violence events in Afghanistan between 1 January 2022 and 30 June 2022. The following steps show how to enhance this bar plot by removing, adding, or editing features.

1. Remove the background gridlines and add thin black axis rulers.

2. Arrange the bars in descending order.

Optional changes

Here are other changes you can apply to reduce the number of visual features to the minimum.

3. Remove vertical axis and vertical axis title.

4. Add labels to each bar with the number they are representing.

5. Consider changing the orientation of the bars.