1.0 Critique of Data Visualisation
The original visualisation can be seen here. 
1.1 Clarity
- The y-axes of the line charts are not synchronised, which might lead to users misinterpreting the pattern and scale of the number of conflicts occurring, especially when comparing large number of events like Battles, versus small, one-off events like Explosions.
- There is no count of events occurring on the plotted map, which leaves to the reader to estimate the counts for each conflict event.
- The dot plot on the map considers the entire time series’ worth of data, which removes the chronological aspect of the data.
- There are no reference points on the line graph to show the count, leaving the user to extrapolate the from the y-axis.
- The sub-events are not plotted, which may be misleading considering that there are sub-events that are “pacifist” in nature, such as “Peaceful Protests”, “Agreement”, or “Change of Group / Activity”.
1.2 Aesthetics
- The chosen colours do not show up distinctively on the map, especially in locations that have many instances of conflicts occurring.
- There is no transparency in the plotted points on the map, leading to overlapping of plotted dots in locations where there are many conflicts happening, for example, in the area depicting Rakhine state in Myanmar.
- The y-axis titles of the line charts were not changed, leaving it as “count of sheet1”, which is not very professional and clean.
2.0 Alternative Design
The proposed design is as follows.
2.1 Clarity
- The y-axes of the line charts are synchronised across categories to improve the visualisation of the scale.
- Different maps are used to view different levels of data, with a pie chart map to visualise the breakdown of conflicts by area, and a proportionate symbol map to show a detailed representation of the various events that occurred.
- Time series can be visualised using a trellis map plot to visualise the number of conflicts that occur over time. Bar charts are used to view a breakdown of all events occurring in the country.
- The total number of events that occurred are shown as values on the dashboard in the detailed proportionate symbol map. This keeps the total events in view even as the user is deep diving into the singular events.
2.2 Aesthetic
- The colours are kept consistent across the maps. Transparency is present for the detailed map to allow some of the overplotted dots to show through.
- The different plots are controlled using a parameter, with the user being able to choose all plots for a quick overview, or different plots to use the correct plot for the correct level of analysis.
- Axis titles were changed to ensure that the graphs remained readable.
- Various parameters are present for Country, State, Event Type, Sub-event type, and Year to allow the user to control their preferences accordingly.
3.0 Proposed Visualisation
Please view the interactive visualisation on Tableau Public here.
4.0 Step-by-Step Guide
4.1 Data Preparation
| No | Step | Action |
|---|---|---|
| 1 | Import the data into Tableau. Change Latitude and Longitude into their respective Geographical Role. |  |
| 2 | Rename Admin1 to State, Location to City, and create a hierarchy. |  |
| 3 | Repeat this for Event Type and Sub Event Type. Reorder both Location and Event hierarchy suitable to their levels. |  |
| 4 | Create a calculated field for Count, putting 1 as the input. |  |
| 5 | Create a parameter called Level, with Country and State om the selection list. |  |
| 6 | Create a calculated field called Lat (Country) with the following formula. Repeat this for Longitude. |  |
| 7 | Repeat the previous step for State Level. |  |
| 8 | Create a calculated field for Longitude Selection with the following formula. This will select the correct longitude aggregation when the respective view is selected. Repeat for Latitude |  |
| 9 | Ensure that the created Latitude and Longitude fields are in the correct geographical role. |  |
| 10 | Create a parameter for view selection for the various graphs and map visualisations. List out the map titles using a custom list. This list will be used for selection later. |  |
| 11 | Create a calculated field called display for the View parameter. This will be used for display selection later. |  |
| 12 | Create a calculated field called AGG to count the number of events that had occurred. |  |
4.2 Pie Chart
| No | Step | Action |
|---|---|---|
| 1 | Drag Long Selection and Lat Selection into the column and row respectively. |  |
| 2 | Change the mark to a pie chart, drag event types into Colour, SUM(Count) into Size and Angle. Add the Display and Count into the Level of Detail mark. Add SUM(Count) into the Text mark. This will create a pie chart on the map visualisation. |  |
| 3 | Add in the relevant parameters in the filters tab. |  |
| 4 | Put the Display calculated field into the filter list. Select the custom value list, type in “Map” and select the “+” option to add this to the list. Repeat for “All”. This will hide the visualisation when other options are selected in the “Select a view” parameter. This will display the visualisation only when Map or All is selected in the “Select a View” parameter. |  |
| 5 | Change colours and opacity to match preferences. |  |
| 6 | The basic pie chart map is created. This allows users to pick and choose the various countries, cumulative events over years, and the breakdown of event type in the country and state level. |  |
4.3 Trellis Maps
| No | Step | Action |
|---|---|---|
| 1 | Place AGG(Full Count), Country and Year in the Level of Detail mark. Place AGG(Full Count) into the Size mark. |  |
| 2 | Place Year and Long Selection into columns, and Lat Selection into Rows. This will create a proportionate geographical map. |  |
| 3 | Add in the relevant parameters for the filter tab. |  |
| 4 | Repeat the step for the Display calculated field, this time using “Trellis” instead of “Map”. |  |
4.4 Detailed Map
| No | Step | Action |
|---|---|---|
| 1 | Change the mark type to a Shape, drag Event Type into Colour, Sub Event Type into Shape, and add Event ID Cnty, Year, Country, & State, into the Level of Detail mark. |  |
| 2 | Drag the Longitude and Latitude into the Columns and Rows respectively. Ensure that they are both in the Dimension format. |  |
| 3 | Add in the relevant parameters for the filter tab. |  |
| 4 | Repeat the step for Display and use Detailed instead of Map |  |
| 5 | Change opacity and colour to suit preferences. |  |
| 6 | The Detailed visualisation will look similar to this. |  |
4.5 Line Chart
| No | Step | Action |
|---|---|---|
| 1 | Create a Breakdown parameter, with All and Country as the custom list for selection. |  |
| 2 | Create another parameter for Event level, with Event and Sub Event as the custom list for selection. |  |
| 3 | Create a calculated field called Event Selection with the following formula to select the view level for events. |  |
| 4 | Create a calculated field called View Breakdown with the following formula to select the breakdown level. |  |
| 5 | Place event selection and YEAR(Year) into columns, and CNT(Event Id Cnty) into Rows. Add another instance of CNT(Event Id Cnty) for a dual axis. |  |
| 6 | Right-click on the second axis, ensure that Synchronize Axis is checked, and ensure that Show Header is unchecked to remove the secondary axis. |  |
| 7 | Change the Level 1 mark to a Line mark, Use View Breakdown for colour, and Event selection for Level of Detail. |  |
| 8 | Change the Level 2 mark to a circle mark. |  |
| 9 | Select Show Mark Labels to ensure that the text is shown on the plotted points. |  |
| 10 | The line chart will look something like this. This will be used to visualise the time series at various levels of aggregation, for All, Country, or State level, & segregated into Event and Sub-event categories. |  |
4.6 Bar Chart
| No | Step | Action |
|---|---|---|
| 1 | Use view breakdown, CNT(Event Id Cnty), and Event Selection into the various marks as seen in the snapshot. Change marks to a bar chart. |  |
| 2 | Move Country into columns, and Event Selection and CNT(Event id Cnty) into rows. |  |
| 3 | Add in the relevant parameters for the filter tab. |  |
| 4 | Repeat the step for Display and use “Bar chart” instead of “Map” |  |
| 5 | The bar chart will look similar to this. This explains the breakdown of events in each country. |  |
4.7 Dashboard Preparation
| No | Step | Action |
|---|---|---|
| 1 | Drag SUM(Count) into the text box. Add in the relevant parameters in the filters tab as required. |  |
| 2 | Format the Label as desired. |  |
| 3 | The sheet will look something like this. |  |
| 4 | Add the Display calculated field, and use “Detailed” for the sheet. This will act as a total count and aggregate numbers in the dashboard and will only appear when the Detailed map is selected. |  |
| 5 | Create a new worksheet, add Event Type to the text box, and repeat the previous steps. This will create a worksheet that will showcase the breakdown of event types. |  |
4.8 Dashboard
| No | Step | Action |
|---|---|---|
| 1 | Drag all relevant sheets into the dashboard. Remove any repeated legends that might appear. |  |
| 2 | Select the down arrow and ensure that the parameter filters are applied to all using the data source. Repeat for any other parameters in the dashboard. |  |
| 3 | The type of charts, level of view, event dates, country, state, event and sub event types can all be changed depending on the information that the user would like to view. |  |
5.0 Derived Insights
- There was a significant spike in violence against civilians in 2016.
The data was further drilled into and viewed at country level. 
A significant proportion of violence against civilians (2,400 events out of 2,700 events) were from the Philippines. 2016 was when current President Duterte was elected. He was elected on a campaign promise to be tough on crime and to eliminate drug traffickers )The Guardian, 2016), creating a climate where anyone can be easily killed in the “war on drugs” (Amnesty International, 2016). Although it seems to be decreasing, there still seems to be significant levels of violence against civilians in the Philippines.
- The country with the greatest number of battles over the past 10 years is in Myanmar.
Drilling down into the data from Myanmar, the largest number of battles happened in Shan State, with 2,040 battles occurring. A similar situation is happening in Kachin state. 
There has been an ongoing ethnic conflict and independence movement in the Shan state against the Myanmar government for almost 50 years (Graceffo,2019). Both Shan and Kachin have armed groups that have allied into a coalition called the Northern Alliance, which has been engaging in various battles with the Myanmar Army (Human Rights Watch, 2019).
- Most battles occur in 3 main areas, Northern Myanmar, Southern Thailand, and Southern Philippines. This is consistent with the various conflicts that are occurring in the ASEAN region. Besides Northern Myanmar that was explained above, both South Thailand and South Philippines have independence movements by the Muslim minorities that reside there.
Drilling down into the time series, disregarding 2020 as it was a pandemic year with many lockdowns in many of the regions, it seems like the number of battles is trending upwards from 2014 to 2019.
By dividing it into country level, we can see that Myanmar and the Philippines have the greatest number of battles, and it may trend further upwards considering recent developments in Myanmar, with a coup d’etat occurring in Feburary 2021. 
There are some missing data in the dataset. The first entries for Indonesia, the Philippines, and Malaysia are 2015, 2016, and 2018 respectively. There seemingly is a downward trend for the Philippines from 2016 to 2019, which may be derived from a decreasing amount of violence against civilians (as shown in point 3).
Most of the explosions have occurred in Thailand, with approximately 56% of explosions over the entire time period, with 2,300 events.
Of this, about 1,600 events occurred in the Southern states of Narathiwat and Pattani.
