Presently, this is the graph presented by a Research Scientist to understand the willingness of the public on Covid-19 vaccination.
Figure 1: Which country is more pro vaccine?
Graph Critiques
| Clarity | Critique |
|---|---|
| 1. | The legend of Vac 1 does not advise what “Vac 1” stands for, and also does not advise what “2”, “3”, “4” stands for. This will mislead viewers from understanding what the bar graphs stand for and how to inteprete the data |
| 2. | The title of the graph on the left does not relate to the questions of the count. What is the population Strongly Agreeing/Strongly Disagreeing to? Are they agreeing to be Pro Vaccine? Or is the population strongly agreeing to take the Covid19 Vaccine once available to them? The title could be clearer in identifying the focus. |
| 3. | The right graph is a repeat of the left graph, but sorted by Strongly Agree Percentages. It does not provide new information from the left, except for an easy visualisation of the sorted countries.It should show the confidence interval of strongly agreed, or the different demographics of the population. |
| 4. | The x-axis of the graph on the right ends before the barchart ends. We are unable to identify what percentage of United Kingdom is. The x-axis should end where the data ends for clear representation of the datapoints. |
| Aesthetics | Critique |
|---|---|
| 1. | The colour of the graph on the left is hard to distinguish clearly, on which is “disagreed” vs “agreed”. It does not give a good visual cue on the separations between the agreed and disagreed crowd. Therefore it does not help viewers easily identify the polarities in each country |
| 2. | The left graph axis has no decimal point, while the right graph has a decimal point on the x axis. It is not consistent. We should keep both axis consistent as they are representing the same data |
| 3. | The name of the countries are all small caps. It should start with Capital Letters for appropriate naming |
| 4. | There is a spelling error in the Title chart of the graph on the left. It should be checked and improved to prevent spelling errors for a more professional look. |
What Can Be Improved?
Figure 2: Sketch of Proposed Design
| No. | Improvements |
|---|---|
| 1. | A scatterplot graph is drawn on the top left to give broad overview on where all countries stand in terms of covid confidence. It plots a score on the Y-axis on scores of the questions chosen, and a score on the x-axis, the willingness of the respondents in taking the vaccine. This scatterplot can help us have a visual exploration on the correlations between 2 questions in the survey and perhaps build some correlation inferences, such as, People are not willing to take the vaccine as they do not trust their government to give an effective vaccine. |
| 2. | A dropdown question bar on the top is created so that we can view different public sentiments with regards to their willingness to take the vaccine, and their confidence and worries about it. The following questions are chosen for the drop down bar. If a Covid-19 vaccine were made available to me this week, I would definitely get it. If I do not get a COVID19 vaccine when it is available, I will regret it. If a Covid-19 vaccine becomes available to me a year from now, I definitely intend to get it. I am worried about getting COVID19. I am worried about potential side effects of a COVID19 vaccine. I believe government health authorities in my country will provide me with an effective COVID19 vaccine |
| 3. | A filter box on the right allows user to toggle between the different countries, and choose which countries they are interested to view. They can view individual countries, or continents, or regions by choosing the Countries of interest. Overcoming the pandemic is also a regional approach. For example, Europe will be unable to recover unless all countries in the EU have the same level of progress and success. This is due to their open border policies that allow for easy movements of people |
| 4. | Various confidence intervals are drawn to depict the sentiments based on different categories, such as Employment Status, Household size, Number of Household Children, Age Group and Gender. We are then able to focus on which group of people requires more convincing in taking the vaccine, and which are ready to take it upon distribution. |
Proposed Tableau Visualisation
Figure 3: Proposed Tableau Design
An interactive experience can be found at this link: https://public.tableau.com/profile/louelle.teo#!/vizhome/Covid19_Sentiments/Dashboard2?publish=yes
Step by Step Guide to Prepare the Tableau Visualisation
Preparing the Dataset
Figure 4: Original Dataset
Dataset is downloaded from Imperial College London YouGov Covid 19 Behaviour Tracker Data Hub hosted at Github. It contains individual files from various countries who have conducted surveys on their population.
Figure 5: Combine Datasets using JMP Pro 15.2
As we are drawing a graph to compare the different countries, we have to combine the files into 1 large dataset. We will use JMP Pro 15.2 to do this as it uses a small amount of computer memory to fulfill this task. This task can also be done with Tableau Prep Builder using the Union feature.
Figure 6: Concatenate Sheets
Add the various country csv files, tick the box Create Source column, and click OK.
Figure 7: Export Combined Dataset
Delete off variables that are not of interest. Keep variables that represent the demographics of the responders, and also Vac_XX variables. Export the JMP Pro file back to excel format by clicking, File, Export.
Figure 8: Export to Microsoft Excel
Choose Microsoft Excel and click Next. Save file.
Figure 9: Add Combined Dataset to Tableau & exclude uninterested variables
Open the Combined Excel Dataset with Tableau 2020.4. The variables chosen are as above.
Figure 10: Update Variables to correct format
Update the variables to the correct format.
- For End Time, update to Date format.
- For Vac 1, update to String format.
- For Source Table, update to Geographic Role, Country/Region and rename to Country.
Creating Parameters for Questions Dropdown
Figure 11: Create Parameter
Right Click Parameters and choose Create Parameter
Figure 12: Parameter Questions
Edit the Parameter as the above to create various questions that the user can choose in an interactive way.
Figure 13: Create Calculated Field for Parameter Questions
Next, we want to link a calculated field to the parameter questions. Click on Analysis, followed by Create Calculated Field
Figure 14: Parameter Codes for the Questions
Create Questions Calculated Field and key in the formula as above. This help links the parameter to the variables.
Update Aliases
Figure 15: Create Aliases
Right click the variables to update the aliases so that when it is plotted the dataset is clearer. Update Vac 1, Questions and Country columns to the aliases as above.
Creating Calculated Fields for Scatterplot
Figure 16: Create Calculated Field of Vac 1 Score
Next, we will create Vac 1 Score using the Create Calculated Field. This score will help update the String format of Vac 1 into a Number format so that we can plot a scatterplot.
To do this, click Analysis followed by Create Calculated Field and type in the formula as above.
Do the same steps as above for Questions to create Question Score
Creating Calculated Fields for 95% Confidence Interval
Figure 17: 95 Confidence Constant
Create Z_95% Calculated Field
Create Number of Records Calculated Field
Figure 18: Only Strongly Agree
Create Only Strongly Agree Calculated Field
Create Proportion Calculated Field
Figure 19: Prop SE
Create Prop_SE Calculated Field
Create Prop_Margin of Error 95% Calculated Field
Figure 20: Proportion Lower Limit
Create Prop_Lower Limit 95% Calculated Field
Create Prop_Upper Limit 95% Calculated Field
Creating Groups
Figure 21: Variable Groups
Right Click Age, followed by Create and Group.
- Group the Age groups according to the picture above.
Continue the same steps for Household Size and Household Children.
Creating Tooltips
Next we will be creating graphs for the tooltips so that this information will appear when a user hovers over the scatterplot.
Figure 22: Tooltip Graph 1
- Add Number of Records to Columns.Update Measure to Sum.
- Add Questions to Rows.
- Add Number of Records to Labels, and update Quick Table Calculation, and Percent of Total.
- Add Questions to Colour.
- Add Country to Filter, and right click Show Filter.
Figure 23: Tooltip Graph 2
- Add Number of Records to Columns.Update Measure to Sum.
- Add Vac 1 to Rows.
- Add Number of Records to Labels, and update Quick Table Calculation, and Percent of Total.
- Add Vac 1 to Colour.
- Add Country to Filter, and right click Show Filter.
Creating the Overview Tab
Figure 24: Overview of Country’s Confidence in Covid19 Vaccine
Drag Questions Score to Rows
Drag Vac 1 Score to Columns
Drag Country to Filter, also to Colour and Labels
Right Click Country on the Filter tab and click Show Filter
Under the drop down bar in Marks, choose Shape.
Right Click on Parameters and choose Show Parameters.
Figure 25: Calculate Average of Vac2 3 Score
Right Click on Vac1 Score and choose Measure, Average
Right Click on Question Score and choose Measure, Average
Figure 26: Adding Tooltips to Overview Tab
Click on Tooltips to insert the tooltips graphs created earlier. Sheets can be added by using the Insert button, and inserting the Tooltip sheets created as above.
Creating Employment Status Strongly Agree 95% Confidence Interval
Figure 27: Bar Chart of Questions vs Employment Status
- Add Prop and Measure Values to Columns
- Drag Employment Status to Rows
- Drag Country to Filters
Figure 28: Questions Format Using Table Calculations
- Right click Measured Values and click Dual Axis
- Right click the x-axis of the graph, and click synchronise axis
- Remove all extra variables in Measure Values till only AGG(Pro_Lower Limit 95%) and AGG(Prop_Upper Limit 95% ) is left.
- Under the AGG(Prop) in the Marks tab, drag Measure Names to Colour
- Under the Measure Values in the Marks tab, drag Measure Names to Path
Figure 29: Add Country Filter
- Right click Country and click Show Filter
Figure 30: Select Multiple Values (List)
Right click the country filter and change the type to Multiple Values(list)
Figure 31: Show Parameter in Legend
Right click Questions and click Show Parameter
Creating Gender Strongly Agree 95% Confidence Interval
Figure 32: Confidence Interval of Questions vs Gender
Repeat the same steps as Employment Confidence Interval for Gender Confidence Interval
Creating Household Size Strongly Agree 95% Confidence Interval
Figure 33: Confidence Interval of Household Size
Repeat the same steps as Employment Confidence Interval for Household Size Confidence Interval
Creating Household children Strongly Agree 95% Confidence Interval
Figure 34: Confidence Interval of Household Children
Repeat the same steps as Employment Confidence Interval for Household Children Confidence Interval
Creating Age Strongly Agree 95% Confidence Interval
Figure 35: Confidence Interval of Age Group
Repeat the same steps as Employment Confidence Interval for Age Group Confidence Interval
Creating Dashboard
Figure 36: Add all Sheets to Dashboard
- Create a Dashboard by clicking on Dashboard on the menu bar, followed by New Dashboard
- Drag all the following sheets to the spaces in the Dashboard
Figure 37: Add Questions Bar to Dashboard
Click on the dropdown bar of the scatterplot to add Parameters, Questions Drag the Questions bar to the top of the dashboard. With that, we are done with the Tableau!
Major Observations
1. Young versus Old
Figure 38: Young versus Old
Older generation of 66 & above, inferred to be the retired population are more pro-vaccine. A higher proportion of them are more agreeable to taking the Covid19 Vaccine, and also have higher trust that their government will provide an effective vaccine. Lastly, they are less worried of the potential side effects of the vaccine.
The younger generation, however, are overall less agreeable to take the Covid19 vaccine.They have less trust in the government providing an effective vaccine, and are more worried of the potential side effect of the vaccine
The most important inference as to why the older generation feels this way, is because for the past 1 year, it is noted that the most vulnerable population to Covid19 are the elderly, and those with pre-existing health conditions.
The biggest casualty of the pandemic tends to be a higher proportionate of elderly population, as their immune system are not as strong and robust as the younger generation. It is also noted that the older population are also more susceptible in catching the virus, as compared to the younger generation. It is seen in the last charts, that the younger generation is not as worried in getting Covid19, as the elderly population.
Therefore we see the trend globally that the elderly population are more open and agreeable to taking the Covid19 vaccine.
2. Large Household with more than 3 Children versus Small Household with 0-2 Children
Figure 39: Large Household with more than 3 Children versus Small Household with 0-2 Children
It is noted that larger household with more than 3 children are more worried about the side effects of the Covid19 vaccine, and are less likely to take the vaccine.
Whereas, smaller household with 0 to 2 children are less worried about the side effects of the Covid19 vaccine, and more likely to take the vaccine.
The conclusion we can gather from this data is quite surprising. As we understand that children under the age of 18 are not going to be vaccinated, they will not be protected from Covid19. However, the parents are most afraid of the potential side effects of the vaccine. The conclusion drawn is that parents may be afraid of side effects the vaccine may cause, that may hinder them from continuing to take care of their children.
The solution to this matter would probably to allow the general population to take the vaccine and offer the vaccine to parents after some time, so that they may witness the efficacy, and safety of the vaccine.
3. Correlations between different countries
Figure 40: Correlations between different countries
From the scatterplot above, we can infer that a countries effectiveness in convincing the population to get vaccinated is strongly correlated to the trust they have in the government.
If a country trusts the government to provide an effective vaccine, they are more likely to get vaccinated. Therefore it is important that governments continue to communicate and be transparent with their nation. This will help build greater trust and confidence in the vaccines.
However, it is also noted that there is a weak relationship between being worried of getting Covid19, and wanting to take the vaccine immediately. Therefore the fear of catching Covid19 will not fuel the confidence in people to take the vaccine.