DataViz Makeover 1

DataViz Makeover 1

Author

Affiliation

Li Zhenglong

 

Published

Jan. 26, 2021

DOI

Original Data Visualisation

1.0 Critiques and Suggestions

1.1 Clarity

S/N Critiques Suggestions
1 The summary write-up described the increase in the proportion of labour force using age group of 55 & over and 25 to 54. However, the age group for the graph is using an interval of 5. Thus, making it hard to link the lead in with the graph. To use a consistent interval of age groups in both graph and text. The dashboard title will serve as an lead-in to summarize the overview insights of the data visualisation.
2 The title “Resident labour force by age - Per Cent” did not effectively communicate clearly if the measure used is by percentage proportion breakdown, participation rate, or difference in percentage change over the years. To reflect the measuring unit LFPR and the comparison between 2009 and 2019. Chart title need to explicitly state the measure and breakdown. “Percentage change in LFPR by Age Group” The comparison period “2009 versus 2019” may be redundant as will be mentioned in the dashboard title.
3 The line chart does not have an explicit Y-axis scale. Without the Y-axis, the figures cannot be ascertained and the measure is not clear.The X-axis can also be clearer with Age group label. Add explicit Y-axis scale and label and X-Axis label in chart title for better clarity.
4 Having only a snapshot comparing 2009 and 2019 can be misleading. We do not know if the changes are consistent between 2009 to 2019. Add sparklines to show a quick trend of each age group category.
5 The line graph may not be effectively used for the categorical data. It may unintended emphasize the change between the age groups instead of the comparison of the two years. Replace with categorical charts using an arrow chart(similar effect as a bar chart highlighting the differences) to emphasize the change in 2009 and 2019.
6 The median age reference line is misleading as the X-axis point is used as the actual age mark in the graph To remove median age and reference line from the chart.

1.2 Aesthetic

S/N Critiques Suggestions
7 The graph shows the redundant year 2009 and 2019 three times. These unnecessary non-data-inks add distraction to the graph. Reduce the number of year labels appropriately.
8 No contrasting colours for comparison. The colour gray and blue are similar and dull and could be hard for viewers with colour blindness. Use contrasting colour-blind friendly colours such as red and blue colours to highlight differences appropriately.
9 Non-data-ink such as the dark colour table and bolding of the X-axis divert focus away from the line chart. Remove dark colour table and un-bold axis label.
10 Too much different information in a single chart diverts the key points. Remove unnecessary chart objects

2.0 Sketch of Proposed Design

  1. Use of simple phrases in title and text to provide relevant and concise information while allowing visualisation to be the main focus and speaks for itself. These wordings allow readers to quickly scan and derive insights.

  2. Reduce non-data-ink by using acronym LFPR and sign % and avoid having too much text to distract or clutter the visual. The Colour of the axis and gridlines will be lightened to further reduce the non-data-ink.

  3. Use intervals of five for Age groups instead of the irregular interval to improve the clarity and avoid skewing the visual to larger groups.

  4. Using Arrow chart allows providing clear percentage growth and the actual LPFR clearly in a creative way. Each arrow represents an age group for their LFPR and changes in 2009 and 2019. This is much easier to see the relative change compared to using bar charts where it may be hard to see when there are too many bars taking up much of the visual space. Or line charts where is used in the original visual may unintentionally emphasize the difference of the age groups instead of the two years.

  5. The sparklines allow quick trending to see what is between the two years without providing too much information diverting the key focus as snapshot comparison can be misleading.

  6. Choice of chart colours will be blue and red which is colour-blind friendly yet simple yet and intuitive to highlight positive and negative

3.0 Data Visualization Step-by-step guide

3.1 Data preparation

3.1.1 Pre-import data prep

Copy the data from Excel sheet “T5_T” in mrsd_2019LabourForce_T5.xlsx into the new sheet Data. Remove the empty columns and format the records as shown.

3.1.1 Import data into Tableau

  1. Import sheet “Data” from Excel into Tableau.
  2. Set field names are in first row.
  3. Select column 2009 to 2019 and right click set pivot.

3.1.2 Rename pivoted columns

  1. Rename column Pivot Field Names to Year
  2. Rename column Pivot Field Values to LFPR %
  3. Format Year as date

3.2 Create Arrow Chart

3.2.1 Drag and drop the dimensions and measure

  1. Drag Age group to Columns
  2. Drag Year to Filters, select only 2009 and 2019.
  3. Drag LFPR % to Rows

  1. Set Marks as line chart
  2. Drag Year to Path

3.2.2 Create Calculated Fields

  1. Analysis -> Create Calculated Fields -> Direction. This is to get the difference of LFPR % of 2019 – 2009 to set the color and shape of the arrow chart. Click on Default Table Calculation to set the definition compute as Year and At the level year of year

  1. Create new Calculated Field, Shape

3.2.3 Create Shape and dual axis

  1. Drag LFPR % to Rows again
  2. Set Marks as Shape
  3. Drag Year to detail
  4. Apply Dual Axis

3.2.4 Create Arrow Head

  1. Create new Calculated Field call shape and set positive or negative based on the [Direction] created earlier.

  1. Drag field Shape into both Color and change the color to red and blue
  2. Drag field Shape again to shape (mark) and change shape palette to filled - triangle up and down for + and - respectively

3.2.5 Remove the extra triangle

i)To remove the extra triangle, create new calculated field call Size of Shape

ii)Drag this new field to Size and this will change the shape sizes are not [Year] = 2019 and set size as 0 which making the triangle disappear.

3.2.6 Complete the Arrow

  1. To change synchronize the line color, select line Mark and drag Shape to Color. Click on Color and set markers to all. This will put the dot to better represent the starting point 2009.
  2. Synchronize the Right Y axis.
  3. Click on Size to increase the arrow line.

3.2.7 Add in the % label

  1. Select Shape Mark and drag Direction to Label
  2. Under label appearance edit label <AGG(Direction)> % to add %.
  3. Under font click match Mark color

  1. Alignment set to Center and Top
  2. Hide repeated values in the Chart by setting the Mark Label “Never Show”.
  3. Remove the Right Y Axis

3.2.7 Title and formatting

  1. Hide all cards on the left to remove the filter/legends
  2. Insert caption for source and X Axis Source: Table 5: RESIDENT LABOUR FORCE PARTICIPATION RATE BY AGE AND SEX, 2009 - 2019 (JUNE)
  3. In format, add grid lines and remove row divider.

3.3 Create Sparklines

3.3.1 Create line chart as base

  1. Drag Year to column, and Age group and LFPR % to Rows
  2. Edit axis to change the Range to Independent axis ranges for each row. So the sparklines will be more obvious.

  1. Sort the Age group by Descending order and hide the Y axis.

  2. From Color set markers to All to show the dots.

  3. Remove the row grid lines and column dividers using format.

  4. Change title to Year-on-Year LFPR to and align center.

  5. Remove x-axis header

  6. Drag Year into label and set mark to label as min/max Remove all point except for bottom age group as shown.

3.4 Combining into Dashboard

3.4.1 Create new dashboard

  1. Drag Arrow graph to the canvas
  2. Drag Sparklines to the right of arrow graph
  3. Drag image object import arrow image[custom image] to bottom right of arrow graph as legend
  4. And drag text objects beside the arrow and key as 2009 and 2019

3.4.2 Annotation, Title size and white space alignment

  1. Add text objects annotation for both charts to highlight the info.

Older age groups show higher increase in %

Younger age groups show more fluctuations in their LFPR

  1. Dashboard title size change to size 15
  2. Chart titles change to size 12
  3. Add blank objects to align the charts.

Data Visualisation makeover completed

4.0 Major observations

  1. Labour force participation rate has increased for all age groups except for aged 20-24 which fall by 2.5% comparing 2019 against 2009.

  2. Older age groups (above aged 55) have a higher increase in LFPR compared to the other age groups.

  1. Year-on-Year for age groups above 30 show gradual changes. However, age groups for 15-19, 20-24 and 25-29, are observed to have more fluctuations in their LFPR from 2009 to 2019.



(https://public.tableau.com/profile/zlli#!/vizhome/LabourForceParticipationRate2009vs2019/Dashboard1)

Footnotes