After more than a year in private preview, Amazon QuickSight is finally generally available in three worldwide regions. Amazon QuickSight is a low cost cloud BI service developed by market leading cloud provider Amazon Web Services (AWS). It is free for a single user and according to current prices ranges between $9 to $24 per user, per month. I can’t resist an announcement like this… time to dig in!
Amazon QuickSight Overview
Amazon QuickSight is a visual front-end cloud BI dashboard solution that also has an API and a data engine. Here is the company’s overview video.
Here is a high level view of the solution architecture.
Amazon QuickSight is part of Amazon’s growing cloud big data analytics portfolio.
Being a cloud-based solution, Amazon QuickSight can scale to thousands of users and be embedded in external applications. Although APIs and embedding are mentioned, I was not able to find them in the docs. On the related community site, I saw that a REST API is planned. I suspect Amazon will rapidly release new features throughout the year just like other cloud BI players in this mega-competitive market.
Amazon QuickSight is currently supported in the following regions:
- US East (N. Virginia) (us-east-1)
- US West (Oregon) (us-west-2)
- EU (Ireland) (eu-west-1)
When you sign up for Amazon QuickSight, you select a home region where SPICE is located. Typically you should pick a region closest to your physical location to reduce cloud BI query latency.
The Amazon QuickSight sign up and getting started process is fast and free. All in all, I was functional in a few minutes.
Slick SPICE Engine
One of Amazon QuickSight’s strengths is a parallel, in-memory, columnar calculation engine called (SPICE). SPICE is capable of handling interactive queries on large datasets. SPICE leverages technology from one of the big data analytics start ups that I mentioned back in 2014 as one to watch – Zoomdata.
One of the more interesting aspects of Amazon QuickSight is that the SPICE engine can be used as a data source for Tableau, Qlik, TIBCO Spotfire and DOMO. An API/ODBC connection is planned for SPICE access from many more BI solutions.
SPICE automatically replicates data for high availability. Each Amazon QuickSight account receives 10 GB of SPICE capacity per paid user, which is allocated when the user signs in for the first time. Free users get 1 GB of SPICE storage.
Amazon QuickSight connects to data in a variety of data sources. You can copy data to the in-memory engine or directly query it with SQL.
When you log in, the cloud BI service automatically discovers your data sources in other AWS services. I noted connectors to data stored on-premises and in the cloud. In order to connect Amazon QuickSight to an Amazon EC2 or on-premises database, you need to add the Amazon QuickSight IP range to the authorized list (whitelist) in your hosted database.
You can upload CSV or Excel files; ingest data from AWS data sources such as Amazon Redshift, Amazon RDS, Amazon Aurora, and Amazon S3; connect to databases like SQL Server, MySQL, and PostgreSQL, in the cloud or on-premises; or connect to SaaS applications like Salesforce. The company shared that data ingestion from Amazon EMR, Amazon DynamoDB, and Amazon Kinesis is planned.
Limited Data Prep
The current version of Amazon QuickSight has limited data prep capabilities. Right now you can manage data fields, change data types and create calculated fields. You can also filter data and perform visual database join operations in a web browser.
For serious data prep, loading, and automating updates, Amazon QuickSight refers the following data integration partners: Informatica, Paxata, Talend, Alooma, and DataBlade.
Basic Data Visualization
Today Amazon QuickSight provides rudimentary data visualizations with some drill-down. All visual types limit the number of data points they display but the amount of points rendered is not known to me. The following chart types are currently available. Most notably I did not see an option for visualizing maps or even adding images.
- Bar Charts
- Line Charts
- Pivot Table
- Scatter Plot
- Tree Map
- Pie Chart
- Heat Map
Another nuance that I noticed during my hands-on evaluation, Amazon QuickSight charts do not contextually filter other charts. You can filter charts by choosing a pop-up menu option or using a filter at the side of the page.
Chart display control properties are extremely limited. You can show/hide legend, refine axis start/end points, customize chart colors and titles. You can also place and size charts anywhere on the canvas as long as they don’t overlap.
Amazon QuickSight includes a built-in suggestion engine that provides recommended visualizations based on the properties of the underlying data. This feature is called AutoGraph. AutoGraph machine learning algorithms learn over time the best visualizations that match your analytical patterns.
AutoGraph seems like a cool capability. I have seen similar features now in other BI tools such as TIBCO Spotfire, upcoming Tableau and other cloud BI offerings. When I tested AutoGraph, it only showed me line charts. I suspect it may still be a work in progress or it will improve over time as I use this solution to train it.
The current version of Amazon QuickSight also includes Storyboards. Storyboards are guided tours through specific views of an analysis. Stories are built by capturing and annotating specific states of analysis. When readers of the story click on an image in the story, they are then taken into the analysis at that point.
Storyboards are similar to what we can do with Tableau or Qlik data storytelling but the depth and user experience between those solutions is vastly different.
To understand differences in visual data discovery capabilities, you do need to evaluate them in a hands-on manner.
For using Amazon Quicksight analytics while on the road, there is a native mobile app for iOS to view dashboards. You can also use any of the modern browsers running on laptops, desktops or a mobile device web browser. A future native Android app is planned along with annotations in an offline mode.
My Initial Impression
The data engine is world class and the documentation was excellent. I was functional in minutes …that does say something. With all due respect to admirable Amazon, what I tested last night did feel immature and awkward from a UX/usability perspective. Like many other data discovery tools that I test, Amazon QuickSight suffers from “BI boxy syndrome“. I have discussed “BI boxy syndrome” at length in a prior BI tool comparison article. User experience makes all the difference.
I am convinced that Tableau’s “Photoshop with data” user experience is the magic that so many other BI tools are missing.
Keep in mind that I am used to working with the top BI and data visualization solutions. I also test a bazillion of them. I am tough to please! For a typical business user, what Amazon QuickSight provides today at a low cost might be good enough…I don’t know. I’d recommend as I usually do…performing a hands-on evaluation and/or waiting for more releases to see how fast it evolves.
While I genuinely appreciate and admire Amazon’s leadership in the cloud and with cloud data sources, Amazon QuickSight’s user experience is lagging. (If anyone from Amazon wants user experience design help from me, my non-compete ends in 129 days. One year non-competes sure feel like forever in a weekly release cloud world.)
Amazon QuickSight’s engine is fabulous but data prep and visualization is lagging behind other cloud BI vendors.
Having said that, I would venture to say Amazon is actually ahead of Google’s Data Studio. I quickly tested it again today just to see how they compared to one another. Google lost that battle but did provide gorgeous templates.
I don’t know why Amazon and Google, two mammoth cloud players with amazing cloud data sources, don’t just buy a smaller cloud BI vendor to expedite entry as a serious cloud BI contender. We all know sizzling front-ends sell the expensive back-end cloud data sources that these two mega-vendors do excel at developing. They both happen to be world leaders in that regard.
That wraps up my quick review of Amazon QuickSight. Although I may have seemed harsh, it was my NOT my intent or tone. I am merely sharing current state with you on this anxiously awaited new cloud BI offering. I look forward to seeing where Amazon takes it and how quickly they progress. I’d bet next year Amazon QuickSight will be far more compelling.
If you want to explore Amazon QuickSight, check out the following resources.
Amazon QuickSight Today
The data engine is world class and the documentation was excellent. I was functional in minutes…that does say something. With all due respect to admirable Amazon, what I tested last night did feel immature and awkward from a UX/usability perspective. It is an early v1 release. I look forward to seeing where Amazon takes it and how quickly they progress. I’d bet next year Amazon QuickSight will be far more compelling.