Thread Closed 
Week 10: World's Biggest Data Breaches
10-30-2014, 04:58 PM
Post: #1
Week 10: World's Biggest Data Breaches
World's Biggest Data Breaches

Link: http://www.informationisbeautiful.net/vi...hes-hacks/

Visualization type : Bubble chart

What is visualized:
The visualization shows world's Biggest Data Breaches among losses greater than 30,000 records in their databases.

Dataset:
Original Dataset is linked at the bottom of the page. Number of records field data is temporal.

Dataset Type:
Table with items aligned to an axis.

Attribute type:
Year of leak-> Ordered ordinal
Method of leak -> Categorical
Number of records stolen -> Ordered Quantitative
Timeline -> Ordered Temporal

Attribute Semantics:
Temporal data since the values varies along the time axis

VisualEncoding Marks and Channels:
BUBBLE COLOUR: Year or Method of leak. (Based on scented widget)
BUBBLE SIZE: Number of records stolen (Based on scented widget)
Yaxis position: Timeline in years
Xaxis position: Data sensitivity or alphabetical (Based on scented widget)

Interactivity:
On double click, user can see the specific news.
On mouse over for a definite time, any smaller bubble is enlarged so that the text is visible.

Filter Scented widgets:
Filter by method of leak and by organization
Order x axis by data sensitivity and alphabetical
Change Encoding for bubble size and color

Pros:
1) Only distinguishable Colors are used. Although there are more hues of red and pink, It was perceivable in my browser.
2) Data ink ratio is higher since it is a bubble chart.
3) Legend changes as we change between scented widgets
4) Due to scented widgets, We can come up with lot of insights and conclusions from the same data.

Cons:
1) "ORDER X-AXIS BY" widget is not so viewable from the visualization.
Since the chart was drawn in first cartesian coordinate, the widget was placed along the x-axis and it is not viewable.
Improvement : But All scented widgets should be placed together for ease of access.

2) Legend for Scented widget Bubble color to encode year was misleading.
Infact, it shows the orange color bubble legend to be interesting stories and there is no legend for the heatmap like coloring based on year.
Improvement : It should have been pointing to interestingness.

3) Filter widget overlaps a part of the graph. The filter widget has transparency to tell us that there is data behind. But when the bubbles are filtered, small and light colored. They are not viewable because of the the filter widget.
Improvment: Move it outside the graph eventhough it takes up more space.

Insight and conclusions from the data through visualization:
1) On the first sight, Adobe and Ebay have largest number of records being stolen over the years.
After scrolling I found Experian to have the largest number of records being stolen.
2) After filter for insider job method of leak, I found easily that Vodafone had data leak due to insider job.


Attached File(s) Image(s)
   
Find all posts by this user
11-01-2014, 12:00 PM
Post: #2
RE: Week 10: World's Biggest Data Breaches
I like this visualization. I think it is designed with exploration in mind. I found just looking at the figure to be a bit overwhelming, all of it doesn't even fit on my screen. The filtering is good. I would like to see the points move closer together after filtering though. When I filter on say "accidentally published" the amount of space taken up by the data is super small but all of the leaks stay in the same place so I have to scroll to find things. I think the graphic could be improved if when a filter is used the data condenses so that I can see it in a smaller space. I also like that by clicking on any one of the breaches it will give you a summary of what happened. I think that the use of color for the year or the hack is redundant and confusing. It took a little while for me to figure out how to read that. The use of color when switched to "method of leak" is very effective though. On the side they have an interactive legend that makes it very easy to read what happened. All in all I think it is a good visualization when used for exploring the data set.
Find all posts by this user
11-01-2014, 12:50 PM
Post: #3
RE: Week 10: World's Biggest Data Breaches
This is an excellent critique, Aravind. I very much agree, especially in regards to your suggestions for improvement.

Another con that I feel worth mentioning pertains to how users interact with individual bubbles. I like the designer's decision to bring the bubble to the foreground with a white outline immediately upon hover over. I consider this to be an elegant and aesthetically pleasing solution. However, I find the delay between initial hover over and enlargement to be frustrating.

After realizing that a bubble can be clicked for more information regarding a breach, a user may want to jump straight to the details for another organization. However, this is not quickly accomplished as users are forced to wait for the bubble to enlarge before the visualization processes their mouse click and shows the details.

A potential improvement for the impatient could be to allow a mouse click to override the enlargement animation and proceed directly to the details as opposed to ignoring mouse clicks before and during enlargement.
Find all posts by this user
11-02-2014, 01:42 PM
Post: #4
RE: Week 10: World's Biggest Data Breaches
This is a very good visualization and critique. It displays lot of data in a single visualization without much clutter. The designer's decision of enlarging the bubble to provide more information is a good idea. The legends and filtering mechanisms provided were very useful in understanding the data, though the unfiltered data can be overwhelming. For the cons, I agree that the delay in enlargement is frustrating and the space occupied can be reduced atleast in the filtered mode. My guess is that the screen size as a whole can be reduced when filtering the data, as the remaining unfiltered data are shown only as very small bubbles.
Find all posts by this user
11-02-2014, 03:34 PM
Post: #5
RE: Week 10: World's Biggest Data Breaches
Quote:2) Data ink ratio is higher since it is a bubble chart.
I don't believe this is really true, there are much better ways to get a higher data/ink ratio than a bubble chart. Although, you could argue that this is much more readable than some of the other methods available for displaying such data. This is a case in which I feel the waste of ink is justified by the ease of readability.

Quote:Cons:
1) "ORDER X-AXIS BY" widget is not so viewable from the visualization.
Since the chart was drawn in first cartesian coordinate, the widget was placed along the x-axis and it is not viewable.
I agree with this, I didn't even notice that this widget existed until I read your critique. That being said, they probably should have thrown the axis controls in some sort of container to make it stand out more.

Quote:3) Filter widget overlaps a part of the graph. The filter widget has transparency to tell us that there is data behind. But when the bubbles are filtered, small and light colored. They are not viewable because of the the filter widget.
Agreed, it is somewhat hard to see data behind the filter widget. I think they probably should have created a separate section above the chart to place all of the filtering/axis controls. (That stands out better than the axis controls.)
Find all posts by this user
11-02-2014, 09:25 PM
Post: #6
RE: Week 10: World's Biggest Data Breaches
I like how the visualization expresses the data and agree on the critique of it. The way the data is being encoded catches attention easily with interesting stories highlighted in brown. Even though the bubbles aren’t encoded quantitatively as far as size to match the data size the viewer still gets an idea of the the larger leaks. I also like the many filtering options provided to help further analyze and compare various sets of data. One thing I though was confusing was the parts were all the small bubbles are crammed together mainly in blue, it is hard to tell the amount of data that was leaked in those scenarios which I can understand can be attributed to space issues.
Find all posts by this user
11-02-2014, 10:40 PM
Post: #7
RE: Week 10: World's Biggest Data Breaches
I think this is a great visualization and its critique by Aravind is excellent. The feature that I like most about this viz is that each bubble is connected to an article about it. It makes the user's experience more complete while exploring the data.

There are a few things that can improve the quality of this viz:
1. Given its size, the user has to do a lot of scrolling to see all the data. By encoding the data sensitivity in color, the user has to either memorize the colors of each category or scroll back and forth to see what category a bubble in the end page falls in.
2. When many bubbles overlap, it is not possible to scroll over to one of the smaller bubbles since its surrounding bubble is enlarged.
Find all posts by this user
11-02-2014, 10:51 PM
Post: #8
RE: Week 10: World's Biggest Data Breaches
Excellent critique by the author. Some of the points I would like to highlight are as follows:
1)The whole visualization does not fit into my computer screen.
2)The data to ink ratio is high in this visualization.
3)The mouse rollover is excellently used in this visualization to show the details about a particular data breach.
4)The filtering widget hides some of the data, making it difficult for the users to see the data.
5)The filtering technique is excellent as it helps the users to concentrate on a particular aspect of the visualization rather than the whole data.
Find all posts by this user
11-02-2014, 11:13 PM (This post was last modified: 11-02-2014 11:14 PM by u0480242.)
Post: #9
RE: Week 10: World's Biggest Data Breaches
Very well done.

Some points:

1. Designer maintains context while focusing on singular item (through mouse click)

2. I thought that the overall context was very hard to take in. I think scrolling up and down was a bit detrimental to the overall idea of the visualization. Limited memory doesn't support remembering what was several scrolls up.

Overall Aravind, I thought that you were very thorough and provided a simple explanation.
Find all posts by this user
11-03-2014, 12:12 AM (This post was last modified: 11-03-2014 12:12 AM by tulasi.)
Post: #10
RE: Week 10: World's Biggest Data Breaches
The visualization is very informative and interactive. I like the enlargement of bubbles when we hover the mouse, however delay is little lagging. We can see more information upon mouse click and its really good idea as we can also navigate to main report from there. The color selections is very nice when I select Method of Leak the way they used Color Legend is visually appealing. It really took some time for me to find out the filed order x axis by and as said by aravind it is hard to find that. I need to scroll page every time and they would have fixed that problem by choosing horizontal view.

I felt the filtering option is good but after filtering there are some small bubbles I don’t understand what they represent, I felt lot of space is because of them and that space would have been used for better representation. When I select No of records stolen button in the top menu bar, it filters data but it is looking like the bubbles are not well ordered in either ascending or descending order. They would have displayed it in a specific order. Other problem I felt was when I filter something using show filter option again I have to click Hide Filter every time to hide that instead there would be better option if it hides when I click anywhere in the visualization. Overall visualization is very effective and showing lot of valuable data but it needs certain improvements as mentioned above.
Find all posts by this user
Thread Closed 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)