Goal the third: Amazon Quicksight

Amazon announced AWS Quicksight at re:Invent 2015 and then released it publicly more than a full year later. After having taking it for a “quick” spin, I’m very bullish on the future of Quicksight (a.k.a. "QS"). It’s pretty simple to get up and going. Maybe not quite as simple as Tableau Desktop, but not too far behind. I had to keep reminding myself that this was the first year of QS, and it’s not going to have feature parity with more mature tools like Tableau…yet. I have every confidence in AWS’s commitment to continued deployment of new features. I also hope they’re looking at more than feature parity. I want to see how the product tackles data visualization differently. Yes, there are some features that are essential, but I want to remain open to discovering new and better ways to complete tasks I do every day.

Accomplishing this third goal with my baseball data was much less time-consuming than my earlier goals. I really just needed to work with some data in QS. I ended up just uploading the full season csv to be loaded into QS. I had already loaded all of the daily files into my S3 bucket, but I could not get QS to successfully access the files. Unable to successfully access the files in S3 was the most frustrating part of the exercise. Nevertheless, here are a couple of screenshots from my analysis:

 Quicksight chart for average fastball spin rate

Quicksight chart for average fastball spin rate

 Quicksight chart showing average effective speeds, by day, for fastballs

Quicksight chart showing average effective speeds, by day, for fastballs

My favorite feature in Quicksight so far? In a standard column chart, I can zoom in and out with the slider (on the side or bottom as shown in the screenshots). I could easily use the filter “calipers” to expand or collapse the current view, or I could move the zoom level across the chart -- that is, move the filter so that still just 100 records are selected as I move across the population. I think this is incredibly useful when I have to deal with a high quantity of values in the chart. I can zoom in just on the top 100 or zoom into a section of the chart that looks a little hinky. The reason this is important to me is that I don’t have to set a new filter on the visualization itself — I'm just creating a window into a section of the chart that I can change on the fly. One small change I would make is to always show the label for the top row in the selection visible. Depending on the window zoom, you might see any one of several different names at the top. I think users should always see the top value in the window.

Result? GOAL COMPLETED

Other posts in this series: