Power BI: Visual of the Week - WordCloud
Analyzing the NASA Facebook page comments
In this blog, I'd like to go over the WordCloud custom visual for Power BI. It's great for showing categorical data for infographics. Since I'm working with a NASA topic, why not have a space background. It makes the WordCloud visual resemble space debris - kind of what Facebook comments are.
The great thing about this custom visual, is that it automatically separates the words in your data set. So if you have a value of "Hello world, my name is Brent." It will separate each word, removing any commas, periods, etc. It also removes common words like conjunctions (and, for, if, etc.). This is great because it removes a lot of the work you would need to do otherwise. Although, you need to turn on the "Default Stop Words" in the formatting tab of the visual to enable this feature.
You can add your own words in this section as well to exclude specific words from the set.
How does it work?
The size of the word is based on the number of times that word appears in the data set. Upper or lower case is ignored. But singular and plurals are separated out.
Further, any apostrophes are split: "I'm" becomes "I" and "m". If you use the default word stop, "I" will be removed in this case. One more good thing to know, entering any specific words into the word stop does not remove the entire sentence from the data set. For example, if I have "Flat Earth", and I want to remove the word "Flat", "Earth" will still be counted in the set.
How is the Performance?
Because it's doing so much work for us, it can be a bit slow to respond. My data set had just over 12,000 comments. Re-adjusting the size dimensions was sluggish, and loading the page took a couple seconds.
There's a performance section with a couple of options. One of them is "Pre-estimate words count to draw". The other is "Quality". I messed around with the pre-estimate, but couldn't really tell a difference in performance with it off or on. The quality % affects the number of results that gets populated in the visual. A higher percentage results in more words on the plot.
Let me know what your experience is with this visual in the comments!
Building an Access Database? Check out my ebook on kindle!