Thread Closed 
Week 14: Pop Music title word frequency
12-04-2014, 09:26 PM (This post was last modified: 12-05-2014 07:56 PM by MitchWest.)
Post: #1
Week 14: Pop Music title word frequency
The visualization I will critique is an interactive visualization from a company called Pitch Interactive who call themselves “A Data Visualization Studio”. I feel like this company should be as credible as any newspaper. This company has done visualizations for companies such as G.E. and Wired Magazine

The vizualization can be found here:

The information this visualization aims to displays is the frequency of words used in Billboard chart song titles.
This visualization uses physical size and location to show the data. Larger words are more frequent and as you go from left to right across a row the words become less frequent. It also using a bit of color saturation espeically in the first few rows. I think this would be called a word doodle or word cloud.
The inteded audience here seems to be music lovers and data junkies. I wonder how many music lovers really care about the frequency of words in a song tttle so it may be the latter .

Tool used:
The visualization was created using processing.

What works?:
Word doodles are intriguing and eye catching. The most common words are clearly the largest and have the most saturation of color.
The color is easy on the eyes and its easy to read the words when zoomed in.

What doesn't work?:
As you get further down the visualization the size and color saturation seems to be the same for all the words. I doubt they all have the same frequency.
We have no idea what the frequency of an individual word is.
There is no way to compare between one decade or another without scanning for a word.
There is no way to filter the data set.

What I would change:

If I were to modify this visualization I would make it interactive. When hovering over a word I would show the number of times that it was used in the time period in the title of the visualization.
I also might include a further breakdown of years or decades of the frequency of that word to break down the data even more.
I would add a way to filter out and search for words. This would allow you to search for words in the different visualizations for easy comparison and to see if the word exists in the dataset.
I would get rid of the color saturation/gradient. This serves no real purpose as you cannot really tell any differences except in the first row of data. Instead I would color the data based on genre. For example if the world was mostly used in rock songs we could make the world black, etc...assuming we could break down the data that way. This would also allow more filtering capabilities.
At minimum I would add the minimum and maxiumum frequency to give some context.
This visualization is not as eye catching as most word doodles because of its linear oreintation.

I do not like this visualization because I do not feel like it conveys any type of information very well. All it really tells us is the most popular words. The rest of the words we cannot differentiate because they all seem the same size.
Find all posts by this user
12-05-2014, 10:04 AM
Post: #2
RE: Pop Music title word frequency
I like the idea of this visualization, but I don't think it's executed very well.

For one, thing I agree with MitchWest that the frequency of words in a song title isn't that interesting to me. What would be interesting to me is the frequency of words in song lyrics (that's what I thought this was at first, and was surprised that "baby" wasn't higher).

I disagree with MitchWest that "This visualization is not as eye catching as most word doodles because of its linear orientation". I actually prefer the linear orientation to the typical randomness of a word cloud, because it's easier to read. However, the creators of this vis did not leave nearly enough space between adjacent words, making it difficult to distinguish between words (especially "I" and "A" towards the top, which look like ,"IA").

I agree that interactivity would greatly improve this vis, especially an option to filter by word to see the frequency of that word (by decade).
Find all posts by this user
12-05-2014, 03:34 PM
Post: #3
RE: Pop Music title word frequency
I like the design critique provided by MitchWest. I agree to most of the points mentioned in the critique. I like the idea of having the interactivity where on hovering the mouse over the word would display the frequency of the word use. However using this visualizations I don't think we are extracting any important information or trend in the data set. "The" is the most frequently used word that's it!!
Find all posts by this user
12-06-2014, 12:56 PM
Post: #4
RE: Week 14: Pop Music title word frequency
Excellent critique MitchWest, I think you are spot on for the most part.

However, I have to agree with aedunn as I prefer the linear orientation as well. While it may not look as pleasing, it does make it much easier to determine ordering.

I like your suggestion for the ability to filter words. One possible way to expand upon this idea is to extend filtering to word types. For example, eliminating all articles or showing only verbs. I think this could make for a much more interesting visualization.

Regarding mentions of being unable to distinguish differences between words in the bottom lines, I believe that this is because there actually are no differences within the bottom lines. Keep in mind, this data is limited to song titles from the Billboard Pop Charts during the given time period. Given that this is a small list and that many songs likely remain in the charts for quite a while, I think it is plausible that we are looking at single occurrences at the bottom portion of the visualization. Zooming in on the 2000 to 2009 visualization (which makes for a great laugh!) was enough to convince me of this as I can think of only one "Top X" pop song title with "Fergalicious" or "Badonkadonk" in it!
Find all posts by this user
12-06-2014, 05:34 PM
Post: #5
RE: Week 14: Pop Music title word frequency
Excellent critique by the author. Some of the points I would like to highlight are as follows:
1) The medium chosen for the visualization is good, as word doodles are always an effective tool for visualization.
2)Color saturation and size are the tools used for showing greater frequency words. These are also effectively used.
3) If the data set was small then, this would have been more effective as after a few words, all the words are of the same size and color saturation.
4) There should have been a better means of user interaction like a pop up over each word showing the number of counts and the time period in which the word occurred.
5) A dynamic filtering mechanism could also have ben added for increased user interactivity and better filtering than only dividing the songs into decades.
Find all posts by this user
12-07-2014, 12:34 AM
Post: #6
RE: Week 14: Pop Music title word frequency
I don't like wordle very much as a visualization tool to actually convey some accurate information, especially this one. Basically all we can capture from this visualization are some high-frequency words, the rest are all background and not really attractive and informative for people who do not really care about it. Besides, as said at the top of that visualization, the biggest words are mostly articles and pronouns, which are kind of less meaningful than with the other, yet they take most of the space and make others less distinguishable. If I were to visualize this dataset, I might just use a table.
Find all posts by this user
12-07-2014, 03:21 PM
Post: #7
RE: Week 14: Pop Music title word frequency
Word doodles are very pleasing to the eye but I did not think the linear orientation was good in this viz because it looks cluttered and like the author of the critique said it does not provide any useful information but only data junk. Color saturation i.e. using the same color for all the words does not help. Adding an element of interactivity showing the frequency of the words could be a salvage point.
Find all posts by this user
12-07-2014, 04:22 PM
Post: #8
RE: Week 14: Pop Music title word frequency
I am not a fan of word doodles because of its chaotic appearance. The only information I get out of this word doodle is that the word "The" is repeated the most followed by all the other words, which is not really helpful or intriguing.

I agree with MitchWest’s critique and the changes proposed also seem quite reasonable, the only exception being that I believe a linear orientation is much better than variations in orientation. Linearity reduces the confusion that the word doodle is so famous for.

All in all, I do not like the visualization, but I do like the critique.
Find all posts by this user
12-07-2014, 07:09 PM (This post was last modified: 12-07-2014 07:10 PM by Nitin Yadav.)
Post: #9
RE: Week 14: Pop Music title word frequency
The visualization here doesn't seem very interesting other than the last bit. Following are the things that I think can be improved upon:
[1] Save on space by not showing the small words that can't be read.
[2] Probably avoid some parts of speech such as: articles, prepositions, conjunctions, etc. because not much information is conveyed from them.
[3] If we could get the frequency count (probably upon a mouse hover) that will help in comparisons.

One thing I really liked about the visualization is the use of color as well as size to indicate the frequency count. I think that makes the objective even more effective.
Find all posts by this user
12-07-2014, 08:47 PM
Post: #10
RE: Week 14: Pop Music title word frequency
I agree with most of the points made in the critique. I don't think the visualization is able to convey any type of information well. I think it would have been nice if common words like "the", "a", and "in" were filtered out. I also think that interactivity would have added an interesting feature and made the visualization better. I think the wordle was a good choice of visualization type. I always think of wordles as visualizations just for fun, and I think the purpose of this visualization is just that.
Find all posts by this user
Thread Closed 

Forum Jump:

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