Talk:Lesson01

A Tour through the Visualization Zoo: A Review

Review 1

‘A Tour through the Visualization Zoo’ article from our per-readings shows amazing ways to visualize and present complex and multivariate data. Interestingly, I am able to relate the details given in this article with some really interesting & interactive graphs I had come across on nytimes.com. They have superbly presented the statistics for the ongoing presidential elections in USA. Some representation types which I could recognize directly were Choropleth Maps and Small Multiples. Most of the other visual representations could be combination of different visualization techniques.

Here are some links for your reference : http://www.nytimes.com/elections/2016/national-results-map http://www.nytimes.com/elections/results/north-carolina http://www.nytimes.com/interactive/2016/03/15/upshot/live-model-republican-primary-results.html

Infovis and Statistical Graphics: Different Goals, Different Looks: A Review

Review 1

I felt that the article "Infovis and Statistical Graphics: Different Goals, Different Looks" is an excellent piece to emphasise the difference between these two seemingly alike graphical depictions but with vastly different goals. I now know how to differentiate between a pure attention-grabbing graphic and a well-developed data visualisation. Good statistical graphics are 1) Interactive (allows user to explore and make comparisons); 2) Integrated (makes data more accessible) and lastly 3) Insightful (improve comprehension and decision making).

As for videos, the 2 TEDx videos showed how the data visualisations could tell a story and they captured my attention, as it should.

One data visualisation which I found interesting and cool is the Norse Attack Map http://map.norsecorp.com/#/

A Reflection on Data Visualisations in the Real World

As I read the articles for lesson 1, I had some flashbacks from my working experience to share.

Robert Kosara's article ("Visualization: It's More than Pictures!") noted that many people still think visualisation is about pretty and colourful pictures, leaving readers to try to figure out what they are trying to tell. I recall in my working experience, some business users tend to like to focus on creating complex and/or many charts and graphs, as if to dissuade others not to read too much into their charts due to the sheer volume and details. I believe some time should be spent on thinking the purpose of each visualisations and does it "pop out" for readers.

The author also argued that examining data is not just about viewing 1 single visualisation, but allowing user interaction to probe the data. I believe this is where tools like Tableau can help fill in the gaps. Business users especially, should spend more time on getting insights from data via visualisations and less time on trying to extract data via complex SQL queries or similar.

However, in reality, many businesses may not have the patience to do data exploration with tools like Tableau, or spend more time than necessary to generate the same charts for a different time period, to be in time for their monthly or quarterly meetings (unless, there are data viz champions like us in the companies!). I believe a middle ground is probably to be able to have simple but effective animations to drive certain points about their charts. For example, I was impressed by the TED video by Hans Rosling ("The best stats you've ever seen"), where simple animations allow the impact of drilling down into data or viewing trends across time. (Note: The tool can be downloaded from www.gapminder.org). It blew my mind because I believe this an effective way of visualisations to attract non-technical folks to understand the charts, via simple animations that make sense and are not fluff. As what Robert Kosara mentions at the end of the article, "Real visualization does not puzzle, it informs".

Yongjian.2015 (talk) 14:21, 21 August 2016 (SGT)Chia Yong Jian

Comment about inclass exercise

The in-class exercise was insightful and Prof. Kam is crafty as always. The reason to pivot the years and values as so that the exploratory graph would be able to be plotted in a line graph and filtered later on. However, that was the hard part, as line graph required a variable that is of type date-time. As such the text format of dates is a problem.

As such the key takeaway I got was how to change the text to a date format using a function in calculated field, and that we need to explore the functions in Tableau further.

Comments about Introduction to Visual Analytics

Thinking back 2 terms ago, evil Prof had introduced to us a form of visual analytics, through graphical exploration in JMP to do data exploration. The first challenge then was “knowing” what to see and how to torture the data to confess the “truth”. The more you torture the data, the more it will confess to you and give you a meaningful story. Hence there is a need to “practice” frequently in torturing the data. By practicing often in data exploration, we will be able to recognise the techniques required to torture the data faster and be able to employ a wider range of tools and techniques.

On the other hand, as the saying goes “A picture is worth a thousand words”. A good and well-designed pictorial representation could showcase the analysis results with conviction to the targeted audience.

Raymond.goh.2015 (talk) 01:52, 22 August 2016 (SGT)Raymond Goh