Today's Viz of the Day on 6/27/2016 is about the deaths in Game of the Thrones. Here is the link to the blog. Lots of context-based graphics made the viz really special and cool. Just love it. The author David Murphy did a lot of research and kept numbers of kills in book, an effort that I can only admire.

In the viz, there is a chart about the cumulative sum of kills since Season 1. I found a few things to tweak.
- The headers are not legible and some are wrapped around.
- Gridlines and column bands are redundant.
- The coloring scheme is redundant
Here are the tweaks I did:
- Changed header and episode names to be short
- The column bands are sufficient in distinguishing the seasons. Gridlines are unnecessary.
- The height of the bars are indicative of the quantity of the kills. Further coloring is not needed.
- Added kills per episode. This gives viewers a better idea on the distribution of the kills.

The resulting chart is:
That's all. The workbook can be downloaded here.

Today (6/22/2016)'s Tableau Viz of the Day presents an overview of the USA presidential election results from 1916 to 2012. This is the 2nd #VOTD in last few days following Matt Chambers's fantastic small multiple design on the same topic, to which I did some tweak to add visual tooltips.

Today's viz is very well designed based on heat maps. Especially, the author Jacob Olsufka has put in great effort in designing the tooltips with various colors, strings, conditional calculations etc. I appreciate deeply how much he cared about using tooltips to convey key information.
The tooltips have a few important numbers about the candidates of democrats, republicans and independents. Here we can further create visualizations for them. Based on an old trick, here is what I got:
It is simple and effective. Click the image to go to the interactive workbook.

The recent Tableau Viz of the Day on spelling bee by Harley Ellenberger is such a delightful masterpiece. It is very sophisticated with carefully selected color palette. I checked every word on it and learnt a few new ones!
Again I found something to tweak. Here is the result:
1.Most gridlines are ditched.
- In the bar chart, the bars of different years are already distinguished by colors. The vertical gridlines don't add any more structure to it at all.
- In the word table, there is no dimension in rows. So, we don't need the horizontal gridlines neither.

2.The jitters chart is replaced by a histogram.
- The jitters give us some vague sense of density. But histogram can give us an exact number per bin. Here the bin is the word difficulty index.
- To create the histogram, we need to turn the difficulty index into a dimension. It is made continuous so that we can add a box plot. Please referred to an early article for some details.
- The box plot is made transparent to be the least intrusive.

Below are pictures on creating the histogram via Index().
3.Have the word list sorted
The author did try to sort the word list by sorting the [Winning word] dimension. Actually we need to sort the [index] in the advanced table calculation settings.
That's about it.

Last week I made my first #MakeoverMonday entry. It is about a data set of Global Temperature 1850-2016. Among all the entries, one was outstanding and won the Tableau Viz of the Day award. It is a simple yet elegant design by Neil Richards, way better than what I did. I downloaded it and tried to animate it, like what I did in my own entry.

As the year started to play, I found something strange. The marks got populated from top to bottom. I immediately realized the year axis (vertical) needs to be reversed. After the reversal, the higher temperatures and later years are on top. After all, we are supposed to raise awareness that the temperature is getting higher visually as year goes by. That's the tweak.

Although there is no right or wrong ordering, we could choose the one that sharpens the point of view we try to communicate to the audience. So there is a more suited ordering than the other depending on the context. A simple design choice may change the degree of impact significantly.

After the tweak: (click images to go to the interactive version)
The original is as follows:

There have been more and more excellent Tableau vizzes out there. There are part of them where I would tweak it in my way. It is far from a makeover though. Just some tweaks. And I would share the changes under a hashtag #TweakThursday. Hopefully, I can contribute some ideas to the community in viz design.

A recent Tableau Viz of the Day on tennis champion Novak Djokovic the King of the World by Ravi Mistry caught my attention. It is excellent in overall design! I like very much the fact that the main color matches the tone of Roland Garros clay courts. Yes, Novak just won his first French Open and completed his first Career Grand Slam (won all four major titles!). What a grandiose achievement!
There are some details I found maybe we can do differently. Below are the main tweaks I did:

1.The tournaments were not sorted nor aligned.

By having them sorted, we can easily count how many Australian Finals he has won or lost etc. The sorting may not be so obvious because I like to cluster the marks together per every round as in the original. I used a continuous pill Index() for the rescue. Open the workbook to see the details.

Also some of the rounds are missing because Novak lost early. We have to make up and hide those vacant data marks.

2.There seem a bit too many gridlines.

Gridlines gave us a sense of structure in the layout. But they are not the only way to show the structure. When the structure is already there, gridlines can become redundant.

For example, in the pivot table, the row bands already give us clear view of every row. So the gridlines are not necessary.

In the line chart, the lines are already separated between years. Thus the vertical lines are in excess.

You might have heard that a major makeover in Tableau 10 is removing the various gridlines on its Desktop interface and canvas. It exhibits a crisper interface with better aesthetics in my view. Note that aesthetics is always something subjective. So, please bear with me.

In the spirit of Tableau 10, I removed most of those gridlines.

3.In the line chart, the three players other than Novak are indistinguishable.

I created a custom color palette with 3 gray colors of different shades to represent the 3 other players. And, Novak is still in that bright orange clay color.

Here is the resulting viz. Click the picture to go to the interactive one.
That's my first #TweakThursday design. Your feedback is welcome. Hope to see alternative designs from others. At the end of the day, my wish is that we can learn from each other and upgrade our skills, much like the #MakeoverMonday exercises by Andy and Andy.

[Update 6/14/2016] 
4.Match data marks with context
This is a tennis story. It looks pertinent to use circle marks, just like the color scheme matching the clay court of Roland Garros.