Author: akalra

For my last blog, I have picked a visualization selected by Bill Gates to be printed in Wired Magazine that he guest edited. He might have his own reasons for choosing this visualization but I see many downsides with this graph.

To start, the audience can infer that the green section representing injuries is significantly smaller than the other two, but it is difficult to judge the relative sizes of the other two sections. Similarly, inside the yellow/pink/green box, it is easy to spot the larger rectangles and get a sense of their relative sizes but again we cannot accurately compare the diseases.

Also, it is easy to read names of diseases in large rectangles but it is straining to the eyes to read inside the small boxes. In addition, few rectangles do not even have a reference label. Even though they appear to be minor causes of untimely death, a designer should not leave out information just for aesthetics of the graph.

Next, I do not understand the need of three different colors. All three colors are segmented similarly in the legend so what is the real need for using too many colors? The same could have been achieved by using just one “stepped” color scheme and separating the three major segments with borders.

Lastly, the 3-D effect doesn’t provide us with any information and on the contrary makes the treemap harder to decode. Another problem induced by this effect involves the darkened colors that appear on the sides of the treemap to represent shadows, which are meaningless and misleading

Solution:
My recommended solution would be displaying the information that appears on the treemap in a simple bar graph. This would convey the story accurately, clearly and would be equally engaging.

References:
Article: https://www.wired.com/2013/11/infoporn-causes-of-death/

Note: Refer article for the visualization

For today’s blog, I have picked up an Info-graphic by TIME which was published close to Women’s Day in 2015. The graph shows how women are represented in politics after 95 years of getting the right to vote.

To some, this visualization might be very engaging but I see many pitfalls in this graph.
Firstly, I feel is the designer targeting the right audience? Is a reader who reads this article just to enjoy pictures with little concern for content and information the appropriate audience for this visualization?
Secondly, the info-graphic shows eight different measures of women’s participation in government and each of the measure is expressed as percentage of female v/s male participation. If they are all same, I do not understand the need to plot these differently.
Thirdly, in the process of making the chart engaging, the designer has exceeded the boundaries of single screen. Information is more powerful when seen together at the same time; this not only saves viewer’s valuable time but also paints complete picture and important connections that may not be visible otherwise.
Fourthly, there is inappropriate choice of media, just to create a variety designer has added pie charts which are a bad choice as already discussed in class.
Lastly, the color choice is misleading. At first glance it makes you think it has something to do with Democrats versus Republicans, while the graph has nothing related to it.

In the end, I feel a simple bar chart with all eight measures would have been an excellent visualization choice. Also, sorting data in order would actually make visualization more meaningful, as the viewer can then judge areas where women representation is best or worst.

References:
http://time.com/4010645/womens-equality-day/
http://www.datarevelations.com/tag/stephen-few
https://www.perceptualedge.com/articles/Whitepapers/Common_Pitfalls.pdf

To study importance of “right scale” let’s see the following graph which is from popular currency exchange website.

Source: www.xe.com

Now suppose you want to know the actual numeric value of right most point, we can see it is little less than half most point between 1.25 and 1.4  (i.e. little half of 1.4-1.25= 0.15), so about 0.6, now adding this to 1.25 it becomes 1.31.

The point I want to convey here is with wrong scaling techniques, it requires more of a mental work than one should actually perform to gain insights from the visualization.  One common source of this problem is algorithm used by common graph rendering software to create these scales. As a designer, one should be aware of this common problem and should consider the following points so that it is easy to perceive values from the graph.

1. All intervals should be equal: This means that the quantitative distance between 2 labels should be equal because if intervals are not equal, it becomes difficult to perceive the values in the graph.
2. Scale should be power of 10 or power of 10 multiplied by 2 or 5: Power of 10 include 10 itself, 10 multiplied by itself any number of times (10*10 or 10*10*10) or 10 divided by itself any number of time (10/10 = 1, 10/100 = 0.1 etc).
Also, it is important to note that 10 multiplied by 2 or 5 is not a constraint in cases where audience thinks of the measure as occurring in groups of any particular size. For example, months (3 or 12), RAM in Gigabytes (4 or 16) etc. A scale of month in form of (0, 5, 10, 15, 20..) is less cognitively fluent than the scale (0, 3, 6, 9, 12..)
3. Scale should be anchored to zero: This does not mean that scale should include zero, instead it means that if scale was to be extended to zero, it should have one of the labels as zero. For instance if we were supposed to extend the above graph the scales in decreasing order would be (0.80, 0.65……….0.20, 0.05, -0.10, -0.25) i.e. this scale has no place for ‘zero’ label hence it is an example of bad scaling.
4. Number of intervals: There is no general rule for this but the scale should provide as many intervals needed for the precision that audience requires but not so many that the scale gets cluttered.
5. Upper and lower bounds of the scale: The general rule is that the scale should extend as little as possible above the highest value and below the lowest value while still respecting the first 3 constraints defined above.
Exceptions to rule 5: a)When using bars, the scale must always include zero, even if it results to an extended scale. b)If zero is within 2 intervals in the data, the scale should include zero.

So next time, it is better to evaluate your scale on these five points before finalizing your graph.
Caution: Above rules apply to only linear quantitative scales.

References:
http://www.perceptualedge.com/blog/?p=2378
http://www.xe.com/currencycharts/?from=USD&to=CAD&view=1D

My this week blog is about how “visual noise” deviates the user from interesting data. Even a credential source like The Economist Magazine falls into trap of beautifying their charts to a level that they lose their purpose.

In their edition “The world in 2012” they published the following chart
The above chart basically matches price of gold to yield of bonds. To somebody who reads The Economist, the above correlation holds substantial value but the visual noise created by distracting image (coin), extremely enlarged chart and microscopic font deviates the attention of the reader.

Following is another such example:
In this chart too, it is difficult to concentrate on the plotted columns while ignoring the cranes and workers that litter the chart. These irrelevant decorations just compel the reader to work harder than they otherwise should to discover the meaning hidden in the data.

A designer should understand that making a chart beautiful to the level that the data looses its integrity actually works against the designer. It makes the chart non effective and fail to provide give quick insights that aid decision making.

The most common chart junk items include:
1. Cartoons or irrelevant decorations: These meaningless decoration do not excite reader about the data rather just add work on user’s side.
2. Dark grid-lines: They often tend to deviate user’s attention. The best practice is to use soft grey grid-lines or eliminate whenever possible.
3. Bright and bold colors: Bright colors are too tiring to look at and also one should be careful about color blind audience.
4. Uppercase: Uppercase should be used only when an element requires special attention.
5. 3-D effects: Three dimensional effect just adds to confusion in readers mind than adding relevant context.

In conclusion, I would state that a good practice after creating a chart is to step back, identify unnecessary items and remove them. Also, one should repeat this process until nothing else can be removed and the visualization has a purpose and supports the objective.

Reference:
http://www.exceluser.com/blog/1133/good-examples-of-bad-charts-chart-junk-from-a-surprising-source.html
https://www.blue-granite.com/blog/data-visualization-remove-chart-clutter-and-focus-on-the-insights

 

While doing my assignment, I was doing some research on how to design a tableau dashboard and what are the key principles for creating an informative dashboard. I found a video explain the Dashboard Best Practices and thought of writing my blog on same.

Building Dashboards involves creativity, science and art and there are 5 key design principles for designing a Tableau dashboard.

1. Have relevant metrics: You need to have relevant metrics for your dashboard which align to the overall strategic goal. A good practice is to involve stakeholders at an early stage to identify the required metrics. Also, it is good to remember, if it doesn’t get measured, it doesn’t get improved; hence make sure that the selected metrics are the ones which can be improved or on which corrective action can be taken.
2. Make it visually pleasing, do not overboard
   The idea of the dashboard is to make it easy for the users to compare and remember data. Take advantage of this but do not go overboard with the charts and try to limit between three to five charts in one frame. Too much information can be confusing and detrimental to the viewer.
3.  Make it interactive:
   Take advantage of the Tableau’s features to create a high level summary of the data but always allow users to explore through the data and get engaged. Give them opportunity to dig to the level of detail to meet their needs.
4. Make it easy to use and access:
   At this point, it is good to consider things like color choices, fonts, layout and also about access, right. Try to answer following questions: Will people be able to click on it, and immediately access it? Will it be fast? Will it run well?
   Focus should be to make a positive experience for the audience and that they can access and use it easily
5. Be open to improvement:
   Be open to improvements and try to collect feedbacks. Creating dashboards should be a continuous process. Metrics and goals might change and a good dashboard should be up to date with those challenges and changes so that it stays relevant.

Keeping in mind the above principles can help us in designing a better dashboard.

Reference: https://www.lynda.com/Tableau-tutorials/Creating-visuals/417094/442256-4.html?autoplay=true

We have already discussed in class, that we should not randomly select your idioms. Also, while doing my last assignment I spent lot of time in “design dilemma”. While figuring out which graph to use, I happened to read an article by Stephen Few in which he mentions about best means to encode quantitative data in graphs. He states that, there is a procedure to follow while creating your visualization.

Step 1: Understand the relationship/message you are trying to present
Step 2: Select the best suitable graph
Step 3: Format your chart

He mentions that almost all typical business information can be addressed by either one or combination of the below mentioned 7 quantitative message types (off course there are exceptions to this) and he has suggested suitable encoding methods which can be a quick cheat guide during our design dilemmas.

Disclaimer: There can be other choices as well, this is just one of the few.
1. Nominal Comparison: When you have to compare between one or more measures in any order.
Suitable Graph: The best encoding method is using either a horizontal or a vertical bar chart, but for large data sets it is better to use simple data points.

2. Ranking: When you have to communicate the order i.e. either highest to lowest or vice versa
Suitable Graph: Again, bar charts are most suitable for this.
Extra tip: For highlighting highest values sort in descending order and for lowest values, sort in ascending.

3. Time Series: When you want to convey how things have changed over time.
Suitable Graph: Line Chart: When you want to stress on the trend and shape of data
Bar Chart: When you want to stress on comparison between individual values
Points + Line chart: To show individual values and simultaneously highlighting shape of the data.

4. Part-to-whole: When you want to represent some values as ratios or part of the whole
Suitable Graph: Bar charts are suitable to represent this relation.
Caution: Do not use pie chart for this, it is difficult to compare size of slices of a pie.
Use stacked bar chart when you want to display both the parts and the whole.

5. Correlation: When you want to compare 2 values and see if there is any relationship between them.
Suitable Graph: Trend line and points (scatter plot) are suitable for this type of relationship.

6. Deviation: To show difference between 2 sets of value
Suitable Graph: Only when displaying time series and deviation together
Line Chart – To stress on shape of data
Points + Line chart – To stress on both on individual values and simultaneously highlighting shape of data

7. Distribution – If you want to measure counts of values per interval along a quantitative scale
Suitable Graph: Histograms are a good fit to emphasize individual values
Use lines to emphasize on shape of data

Reference: https://www.perceptualedge.com/articles/Whitepapers/Communicating_Numbers.pdf

The following graph is from Bloomberg (2013); which for many is a trusted source. Unfortunately, even this trusted source has misused power of statistics to deceive people.

Looking at this graph, a common man would be highly concerned with the slope depicting sharp decline in median income for U.S. men but in true sense there are more flaws with the graph than with the fact depicted.
The first flaw is regarding incomplete information. The designer has only shown 2 data points and no information is depicted about what happened in middle years. On investigating more from U.S Census data one can see that median income was actually stable between 1972 and 1999 which is contrary to what designer has depicted. Also, for age 45-54 there was actually an increase in median income till 2000 and only after that there was a decline in the income.

The second flaw is with the y-axis. The designer has deliberated truncated the y-axis so as to magnify the gap. If the same graph is seen making y-axis start from zero, the decline doesn’t feel much and our perspective about the problem changes.

Lastly on investigating on the data more, we find that from 1947 to 1972 there was steady increase in median income and since 1972 (end of Gold standard) there has been a slow decline in the number. The designer has deliberately chosen 1972 and 2012 to catch attention of its readers. The same news can be changed to “Income for men has risen” by giving 1947 and 2012 as new data points.

References:
Image & Article Source: https://www.bloomberg.com/news/articles/2013-12-31/for-u-s-men-40-years-of-falling-income
Other Source: https://medium.com/i-data/misleading-with-statistics-c63780efa928#.qaw475rwg

The following graph was posted in 2011 and tried to use proportionally sized circles to depict the visualization.

If we closely observe the visualization, we can spot many pitfalls in the used media.

• The first flaw is with the size of the circle; the size does not match the associated data values and hence exaggerates the small amount of money donated and number of deaths caused by each disease. This is a very common mistake which designers make while using circles to show their graph. Various design software only allow height and width adjustment and designers often fall into trap of adjusting diameter of the circle rather than area to match their data.
• The other flaw is with choosing the inappropriate display media. It’s difficult for one to track different colors and then match information on the legend accordingly. The image involves 3 way process of first looking at circle in “Money Raised” area; then mapping the color and finding details in the legend and finally looking for the same color in “Deaths” area. It would have been easier for the reader if the name of the disease and respective money raised and deaths were placed together.
• Third issue is with use of colors. As discussed in our class, a good number of people are color blind and therefore it is not a good idea to use too many colors. Colors are also an issue when somebody wants to take a handout of your visualization for their reference. Also, as the size of last few circles are very small, it is difficult to spot the color and map it with the colors in the legend.

References:

Image Source: http://cdn3.vox-cdn.com/uploads/chorus_asset/file/663618/Donating.vs.Death-Graph.0.jpg

Other Source: http://www.huffingtonpost.com/randy-krum/false-visualizations-when_b_5736106.html

 

                      Source: © Understandingusa (2007).

The above picture was published in www.understandingusa.com in 2007 and represents Civilian War Casualties in 20^th Century by country.
The designer had idea of a “shooting target” as a reference behind the above visualization and as maximum points are earned in the center, the innermost circle represents the countries with maximum casualties.
The flaw with above visualization is related to “how different people perceive same thing”. In this case, the problem is, not all people are aware of shooting sports and a lay man would correlate this with concentric circles and without looking at the description would infer the biggest circle to be countries with maximum number of deaths, which is not true. Only, reading the graph carefully with description would make it clear that such is not the case.
The other issue is regarding choice of image; image of a family is used as a target which might induce emotional harm to some of the viewers. Again the issue is related to perception, this graph might be acceptable for people who regularly use such visualization to gain attention of audiences but might be inappropriate for people less inclined towards such crude representations. Therefore, it is always better to consider various perceptions and alternative methods of raising awareness than using such arousing graphics before designing a graph.

 

Reference:

• Sabrina Bresciani. 2008. Disadvantages Associated with Graphic Representations of Information . [ONLINE] Available at: https://pdfs.semanticscholar.org/23d2/3f5152c9b8b34f104b43d1c862ee62d2edac.pdf. [Accessed 17 January 2017]
• SAGE Journals: Your gateway to world-class journal research. 2017. SAGE Journals: Your gateway to world-class journal research. [ONLINE] Available at: http://journals.sagepub.com/doi/figure/10.1177/2158244015611451?. [Accessed 22 January 2017].