Original Data Visualization
The original visualizations displays data from The Armed Conflict Location & Event Data Project, which collects information on all reported political violence and protest events around the world. Data can be extracted from the following link. A screenshot of the data visualization that will be madeover is provied below:
Part A: Critic of the graph for its clarity, aesthetics and interactivity
Clarity
1. Scale of y-axes of graphs is not consistent: The y-axes of the line graphs on the right of the visualization are not adjusted to the same scale, which impedes comparability of information across the line graphs. As seen in Figure 1, the scale of ‘Riots’ is up to 20 whereas the scale of ‘Battles’ is up to 700. Even though the disparity is big, the line graphs do not show this disparity due to the lack of a consistent scale. As seen in Figure 1, in year 2018, the user is misled into interpreting that the armed conflict event type with the highest count is ‘Riots’ as it has the highest peak, but the count of ‘Riots’ is actually the lowest compared to the other event types.
2. Labels of y-axis on the line graphs is unclear: The line graphs in the visualization are labelled as ‘Count of Sheet1’. As it is unclear what ‘Sheet1’ refers to, the y-axes can be more clearly labelled to represent the information that the data points illustrate.
Aesthetics
1. Good usage of color legend to represent different event types: Points on the map are represented in different colours based on their event type, making it easy for the user to differentiate points on the map by their event type.
2. Lack of consistency in colors of the legend across graphs in the dashboard: In the map, a color legend is used to indicate the event type, where blue is used to represent the event type ‘Battles’. However, in the line graph, the same blue is used to represent all event types.
Interactivity
1. Tooltips on the map gives code that users cannot directly interpret from the graph and are meaningless to the user: Ideally, a visualization should be able to stand on its own, allowing the user to fully understand the visualization without using an external codebook. However, the tooltip over the map contains ‘Latitude’ and ‘Longitude’, which require a user to refer to a user would need to refer to a codebook to interpret the name of the town where the event took place.
2. Redundant information is provided through the tooltip: The tooltip over the map contains ‘Event ID Cnty’ which is redundant as these ID numbers are arbitrary and are generally for the purpose of database management. The ID is thus not useful to the user and should not be provided in the tooltip. Another case of redundancy is the tooltip over the line graph which contains ‘Year of Event Date’. As users can clearly interpret from the x-axis of the graph which year each data point refers to, supplementing the information on ‘Year of Event Date’ in the tooltip is redundant.
Part B: Suggesting an improved alternative graphical presentation
With respect to the earlier critics, the following suggestions have been made:
Clarity
1. Combining lines onto the same graph: Instead of using different rows for each event type, I will be combining all event types onto the same graph differentiated by colour. This will enhance compariability between event types of a country.
2. Improvement of graph labelling: To avoid ambiguity, all graphs will be labelled clearly.
Aesthetics
1. The same colour legend will be used across graphs in the visualization: A consistent color scheme differentiating event type will be used across graphs in the same visualization where applicable.
Interactivity
1. Provide information that is meaningful to the user: Instead of providing ‘Latitude’ and ‘Longitude’, I would provide the name of the town to the user.
2. Remove redundant information provided in the tooltip: The tooltip should be concise, providing information that adds value to the user’s interpretation of the visualization. Hence, both the ‘Event ID Cnty’ and ‘Year of Event Date’ from the tooltip.
In addition, the following enhancements will be made:
1. Map for summary statistics: Another map will be added to provide aggregate summary statistics for every country. This gives the user higher level information on the severity of political violence and protests in the region.
2. Provision of more useful information in the tooltip: Besides the corrections mentioned above for interactivity, additional descriptors will be provided in the tooltip.
The proposed alternative data visualization is as follows:
As seen in Figure 2 above, the final dashboard contains 3 visualizations - 2 maps and 1 line graph. The maps help to provide a regional and country overview of the political violence and protests, whereas the line graphs provides greater insight into the progression of these events over time for a country.
Part C: Designing the proposed data visualization using Tableau
The visualization is available in the following here.
Part D: Description of Data Visualization Preparation
1 Data Preparation
After extracting the data for South-East Asia into a .csv file, we must filter the data for years 2015 to 2020, which can be done in Excel before importing the dataset into Tableau. Sorting the rows of the dataset in descending order according to the column Year would place the rows for years 2010 to 2014 at the bottom of the dataset. These rows can be selected and deleted so that only years 2015 to 2020 remain. Subsequently, the dataset can be imported into Tableau.
2 Creating the Worksheets
2.1 Creating the symbol map for country overview
Create a new worksheet by clicking on New Worksheet at the bottom pane, and name it as Country Map.
2.1.1 Adding Elements to the Map
- Ensure that Latitude and Longitude have been assigned these respective geographic roles. Verify this by clicking on the down arrow on Latitude and Longitude in the left pane > Geographic Role > See that Latitude and Longitude have been checked.
- Add Latitude to column and Longitude to row
- Assign both Latitude and Longitude as Dimension
- Click on Show Me at the top right corner, and select on Symbol Maps
- Drag and drop Event Type to Color
2.1.2 Inserting Filter for Country
- Drag and drop Country to Filters, and select all countries to be included in the filter
- Click on the down arrow on the filter > Show Filter.
- At the filter in the right pane, click on the down arrow at the top right corner of the filter and select Single Value (List).
- Click on the down arrow at the top right corner of the filter again > select Customize > uncheck Show ‘All’ Value
2.1.3 Formatting the Title
- Double-click on the Title to rename it
- Rename the title as ‘Armed Conflict Location and Event in
, 2015 to 2020’ as shown in the figure below:
2.1.4 Formatting the Tooltip
- To format the tooltip, double-click on tooltip.
- *Removing Latitude and Longitude**: Delete latitude and longitude from the tooltip so that Event Type is the remainin row in tooltip.
- Drag and drop Location, Sub Event Type, Event Date, Fatalities and Actor 1 to Detail. For Event Date, select Date as the format.
- Double-click on tooltip and enter in every dimension as follows:
2.1.5 Screenshot of the completed worksheet
With that, the worksheet has been completed. A screenshot of the worksheet is below.
2.2 Creating the map for regional overview
Create a new worksheet by clicking on New Worksheet at the bottom pane, and name it as Region Map.
2.2.1 Create Calculated Fields for Every Event Type
For the tooltip to display the count of every event type, calculated fields can be utilized. - To create a calculated field, select the down arrow on the top of the left pane > Create calculated field… - Name the calculated field as Battles with the formula in
- The same process can be repeated for Explosions/Remote violence, Protests, Riots, Strategic Developments and Violence against civilians. To speed up the process, duplicate the calculated fields and simply rename the fields where required. To duplicate a calculated field, click on the down arrow on the field in the left pane > Duplicate. Then click on the down arrow on the duplicated field > Edit.
2.2.2 Adding Elements to the Map
- Ensure that the Geographic Role of Country has been assigned to Country. Verify this by clicking on the down arrow on Country in the left pane > Geographic Role > See that Country has been checked.
- Drag and drop Country to column
- Click on Show Me in the right pane > Maps
- Drag and drop Event Type to Color. Select on Event Type > Measure > Count. To format the color scale, click on the down arrow on CNT(Event Type) in the right pane > Edit colors. I selected Orange-Gold as the color scheme to match the topic.
- Drag and drop the calculated field for every event type into Detail (i.e. Battles, Explosions/Remote violence, Protests, Riots, Strategic Developments and Violence against civilians).
- Drag and drop Fatalities to Detail.
2.2.3 Formatting tooltip
- Double-click on tooltip to open up the editing screen, and edit the tooltip as follows:
2.2.4 Formatting title of the worksheet
- Double-click on the title of the worksheet > Rename the title as ‘Armed Conflict Location & Event in South-East Asia, 2015 - 2020’.
2.2.5 Screenshot of the completed worksheet
With that, the worksheet has been completed. A screenshot of the worksheet is below.
2.3 Creating the line graph
Create a new worksheet by clicking on New Worksheet at the bottom pane, and name it as Line Graph.
2.3.1 Create Calculated Fields for Every Event Type
For some countries, the event type can be 0 in some years. If we were to plot the event type, the line graph will be discontinuous, such as in the figure below. Thus, a calculated field needs to be created to insert a value of 0 for the years where there is 0 count for an event type, so that a continuous line graph can be plotted.
- To create a calculated field, select the down arrow on the top of the left pane > Create calculated field…
- Name the calculated field as Count_Event Type with the formula in
2.3.2 Adding Elements to the Graph
- Drag and drop Event Date to column. Ensure that year is chosen, such that the column appears as YEAR(Event Date). If not, click on the down arrow on Event Date > Year.
- Drag and drop Count_Event Type to rows.
- Drag and drop Event Type to Color.
2.3.3 Inserting Filter for Country
- Drag and drop Country to Filters, and select all countries to be included in the filter
- Click on the down arrow on the filter > Show Filter.
- At the filter in the right pane, click on the down arrow at the top right corner of the filter and select Single Value (List).
- Click on the down arrow at the top right corner of the filter again > select Customize > uncheck Show ‘All’ Value
2.3.4 Formatting the Graph
- Select Color > Markers > All
- Rename the title of the worksheet as ‘Armed Conflict Event by Type’
- Right click on the y-axis > Edit axis to rename the axis as Total Events
- Right click on the x-axis > Hide field label
- Double-click on tooltip
2.3.5 Screenshot of the completed worksheet
With that, the worksheet has been completed. A screenshot of the worksheet is below.
3 Combining the Sheets into a Dashboard
- Create a new dashboard by clicking on New Dasbhoard at the bottom pane, and name it as Dashboard.
- In the left pane, Click on the down arrow next to Fixed Size to change the setting to Automatic.
- Drag and drop Region Map onto the dashboard, add Line Graph below it and then add Country Map to the left.
- The dashboard should contain 2 filters for Country. Under the filter for Country for Country Map, click on More Options > Apply to worksheets > selected worksheets > Check Line Graph. This would enable the filter to be applied to both the map and the line graph. Thereafter, we can remove the other filter for Country.
- For the color scale for Region Map, click on More Options > Edit Title and rename the title as Total Events.
- Resize the elements within the dashboard until the final dashboard is derived as in the Tableau link.
Part E: Major Observations from Visualization
From 2015 to 2020, the total armed conflict events that happened in a country was highest for the Philippines and lowest for Laos: From the Region map, based on the color scale, we can infer that the total events is highest in the Philippines which has the deepest shade whereas Laos has the lowest number of events as it has the lightest shade.
From 2015 to 2020, the Philippines had the highest fatalities to armed conflict ratio: Observing the summary statistics in the Region map, Philippines stands out for its high fatalities to armed conflict ratio. Every other country had a fatality count that is lower than the armed conflict count. However, Philippines had a higher fatalities count than armed conflict count. A screenshot of the statistics for Philippines is given below:
- From 2015 to 2020, events in most countries were distributed relatively evenly throughout the country. The exceptions are Thailand and Myanmar: From the Country map, most countries did not have a distinct trend in the distribution of data points for every event type. Events by battle type were relatively evenly distributed throughout the country. The first exception is Thailand, which had a very high concentration of Battles in the Southern provinces, namely Narathiwat and Pattani. A screenshot of Thailand is shown below:
The second exception is Myanmar, where it can be observed that Battles are concentrated in the Western and Eastern sides of the country while Protests are concentrated in the central to Southern parts of the country. Battles are concentrated in the provinces of Shan and Rakhine, whereas Protests are concentrated in provinces like Yangon, Ayeyarwady, Magway and Mandalay. A screenshot of Myanmar is provided below.
- Protests is the most common most frequent event: From the line graph, Cambodia, Indonesia, Malaysia and Vietnam had Protests as their most frequent event. This is evidenced by how the red line graph is higher than the other line graphs, such as in Cambodia’s case.
- Philippines and Vietnam saw a clear declining trend in events from 2015 to 2020, while Indonesia and Thailand saw a clear increase in events from 2015 to 2020, such as Indonesia and Thailand: The line graph allows us to interpret changes in trend over time.Many of these trends had ups and downs. Putting aside the trends that were a very low proportion of the total events of a country and focusing on the key event types, I noticed that some countries exhibited clear increasing trends while others exhibited clear decreasing trends. Other countries saw some improvements and some deprovements throughout the 5 years, such as Cambodia and Myanmar.It is worth noting that the increase in events for Thailand is likely due to the 2020-2021 protests.