Big data just keeps getting bigger — we generate 2.5 quintillion more bytes every day. Many enterprises use business intelligence (BI) software to find the signal in all that noise. Visualizing data through BI tools allows users to process information quickly — according to MIT, human brains can process images in as little as 13 milliseconds.
QuickSight is Amazon's entry in the BI space. Let's take a look at what Amazon QuickSight is, how to use it, and some alternative options.
Amazon QuickSight, a component of Amazon Web Services (AWS), is a cloud-based serverless business intelligence platform that allows users to create visualizations and dashboards. QuickSight takes advantage of machine learning to identify anomalies in data and make predictions through its ML Insights feature.
QuickSight supports a variety of data sources, such as individual databases (Amazon Aurora, MariaDB, and Microsoft SQL Server), data warehouses (Amazon Redshift and Snowflake), and SaaS sources (Adobe Analytics, GitHub, and Salesforce). It also supports files in ELF/CLF, CSV/TSV, and XLSX formats, as well as semistructured files in JSON format.
QuickSight comes in standard and enterprise editions. The standard edition, which doesn't include ML Insights, costs $9 per month for an annual subscription and $12 for a monthly subscription. Those rates apply whether users are creating or just viewing visualizations.
The enterprise edition offers advanced features, such as encryption at rest and support for Microsoft Active Directory, and it allows readers to take advantage of QuickSight's pay-per-session pricing. Users of the enterprise edition are classified as authors or readers based on whether they want to create or simply consume visualizations. Readers pay just $0.30 for a 30-minute session. Sessions are renewed automatically every 30 minutes and time out after 30 minutes of inactivity. Authors pay $18 per month for an annual subscription and $24 for a month-to-month subscription.
To start creating a visualization, log in to the AWS Management Console, search for Amazon QuickSight, and create a QuickSight account. Choose a subscription type (standard or enterprise edition), pricing plan, which region your data is stored in, and whether automatic discovery of data from other AWS services is allowed.
Once your QuickSight account is ready, kick off a new analysis by getting data into the tool. You can create a new dataset by either extracting information from an existing data source or uploading a file.
You can directly query databases, but for other data sources, QuickSight relies on SPICE — Super-fast, Parallel, In-memory Calculation Engine — to handle massive datasets and deliver results quickly. You can also import datasets from databases into SPICE to improve performance.
Every Amazon QuickSight account receives 10GB of SPICE capacity for a paid user and 1 GB for a free user, and you can pay to increase your account's capacity up to 1TB, with the option of requesting capacity over this limit if necessary. QuickSight's admin page shows you how much SPICE capacity you have remaining, and lets you delete datasets loaded into SPICE to free space for analyzing other data.
After data is loaded into the tool, you can prepare it for analysis. As part of data preparation, you can delete unnecessary columns, add new calculated fields derived from existing columns, apply data filters, format columns, or change field names and data types. You can save prepared datasets for use with multiple analyses.
After data preparation, you're ready to create a visual. QuickSight supports more than a dozen visualization types, including bar charts, pie charts, pivot tables, and heat maps. If you can't decide which type best showcases your data, QuickSight's AutoGraph feature can automatically choose an appropriate type.
You can customize visualizations by altering fields, rewriting titles, and changing visual elements with drag-and-drop functionality.
You can combine visualizations into an analysis — QuickSight's term for a group of visuals or stories. Each analysis can contain up to 20 data sets and 20 visuals, and you can connect and present multiple iterations of an analysis together to tell a business story.
You can use analyses to create a dashboard — a read-only snapshot used for reporting data via a web browser or an Android or iOS app. Users can share dashboards and datasets with other QuickSight users and groups or publish a dashboard to send a link to all subscribers of the dashboard by email. You can also email links to analyses to other QuickSight users in your account.
Tableau has been around since 2003 and features both desktop and cloud-based platforms. Gartner has rated the product as a leader in the Magic Quadrant for Analytics and BI Platforms seven years in a row. Tableau features more visualization types than QuickSight, and it also has connectors to a wider range of data sources than QuickSight, including on-premises databases.
Power BI, released in 2013, is a cloud-based platform that allows both technical and nontechnical users to create reports, dashboards, and visualizations. Like Tableau, it features integrations with on-premises databases, but it requires the use of Azure — Microsoft's cloud computing platform — to store data in excess of 10GB. Power BI is less expensive than Tableau, has a more robust set of data visualization features than QuickSight, and is accessible to nontechnical users.
Google Data Studio, released in 2018, offers streamlined data visualization and dashboarding. It has more connectors than QuickSight and a more intuitive interface. Data Studio is best suited for enterprises that require a lightweight BI product for nontechnical users, especially if the organization is already in the Google ecosystem.
Stitch can help replicate all your data into a warehouse for use with QuickSight or any other BI tool. It supports data replication from more than 100 data sources, including databases like MariaDB and MySQL and SaaS tools like GitHub and Salesforce. Sign up now to try Stitch for free.