Business Intelligence – BI tools such as AWS QuickSight, Tableau, Power BI, IBM Cognos (and many more) are designed to assist companies in generating business insights with the help of visuals. Since BI market is highly competitive, therefore, the companies that have provided these solutions have added distinct features in order to target a certain customer base that might use those features in their business requirements.
In this article, we shall make a brief comparison between the two Business Intelligence solutions: AWS QuickSight and Microsoft Power BI. We will first talk about the two technologies separately, highlighting the key features that each tool provides, and also the pros and cons of using the two. AWS QuickSight is a cloud-based BI solution (which runs on Amazon Web Services Platform) that you can use to build visuals, perform ad-hoc analysis, generate business insights, and share the results with others. AWS QuickSight connects to a variety of data sources, including AWS data (S3, Athena, Redshift etc.), third-party data, spreadsheet data, and more. QuickSight processes the data through SPICE, which stands for Super-fast, Parallel, In-memory Calculation Engine. Amazon claims that it is a robust in-memory engine that performs advanced calculations and serve data. If you want to create a dataset in QuickSight, you can either import it into SPICE, or perform a direct query (which is a way of directly querying the data instead of importing it into the tool). It is recommended to use SPICE to load the data so that QuickSight would be able to access it quickly and efficiently. On the other hand, direct query accesses the data by querying the source data directly, but this method is considered inefficient in QuickSight because data is queried every time a change is made in the analysis.
Here is a high-level architecture of QuickSight. As a summary, Data source connects to SPICE, which loads and processes (data cleaning, transformations etc.) the data. This data is then fed into QuickSight for data visualization. The visual presentation of QuickSight is one of its key selling points. Although the quantity of visual types might not be there, but for QuickSight, it’s about how it appeals to the end-user. Following are some of the major components/key features in AWS QuickSight:
- Visuals – These are the components you use to represent your data in the form of visuals. You have bar charts, box plots, combo charts, heat maps, KPIs, line charts, and many more visual types that you can use to create meaningful reports.
- Insights – As the name suggests, this feature allows you to generate insights with the help of built-in machine learning algorithms. This feature is quite useful because it allows you to interpret your data in a way that might add value to your analysis.
- Sheets – These are like separate pages that you see in Power BI, where you can keep a group of visuals on a single page. You can have one sheet showing visuals that represent the sales of a company, and another page showing visuals related to inventory analysis.
- Simplicity – Although this is not a proper ‘feature’, it is something certainly important for this BI tool to appeal to the market. Even people without much technical knowledge can easily explore data and extract valuable insights because of the simplistic and intuitive nature of the tool. You are most of the times performing simple operations (related to arithmetic, string, dates etc.) on the data, data type changes, and dragging and dropping fields on to the visuals.
- Speed – Speed is a major selling point for QuickSight, due to its SPICE Engine.
]Using all of the features provided by QuickSight, users can create meaningful, beautiful, and interactive reports to assist stakeholders in various business areas, such as:
- “Pros and Cons”
- Just like any other tool, QuickSight has its pros and cons. Here are some of the pros of using this tool.[/vc_column_text][vc_column_text]
- Easy to Use – As mentioned earlier, QuickSight is very simple and intuitive to use. The users can configure and start using the tool in no time. It also takes less time to learn the tool, so if this is your first BI tool, working on the data and creating visuals would seem very easy.
- Everything is on the Cloud – As QuickSight runs on the AWS platform, you don’t really need to set up anything on your system. You would just need a working AWS account, a subscription, and a network connection so that you can easily access the tool via web. Even if you have a low-end system, Quicksight would run flawlessly since everything is hosted on the AWS cloud platform. You can also access the tool via Android or IOS device, since the integration on mobile devices is also excellent, and would allow the users to view the content in a seamless manner.
- Quality of the Visuals – QuickSight has some stunning visual types in its collection. Although they are limited in quantity, they can certainly make a huge difference in visual presentation.
- Pricing – QuickSights pricing is pretty optimal for an average user, compared to other platforms. For more information on pricing, please visit the link here.
- Speed – This is one of the key features of using QuickSight. Because of the SPICE engine, data loading and processing becomes a great experience for all levels of users.Now that we have highlighted the pros, here are some of the cons of using QuickSight.[/vc_column_text][vc_column_text]
- Limited Visual Types – As mentioned earlier, QuickSight has quality visuals, but they are limited in quantity. So, if you need a visual that is not present in the collection, you might need to look for an alternative within the available visuals set.
- Simplicity – The ease of use and simplicity was highlighted as an advantage, but it’s also one of its major disadvantages. Now this totally depends upon the use-case. If your reports require simple data connectivity, simple calculations, and visuals that only need fields to be dragged and dropped on to them, then QuickSight is a great choice. But for cases where we have to perform high level transformations, calculations, and complex reporting, this tool is not an optimal one. For complex reporting, there are tools such as Power BI, which we shall talk about in the next section.
- Still New to the Scene – Quicksight is still pretty new in the BI market. So, this solution has to play catch-up with other competitors such as Tableau, Power BI, in terms of adding new features which would support complex data processing, reporting, or sharing, so that it appeals to the mass market and in particular, the big corporations.
Now that we have talked briefly about QuickSight, let’s take a look at its competitor: Microsoft Power BI.
Microsoft Power BI: As the name suggests, Power BI is a Business Intelligence software product created by Microsoft. It combines business analytics, data visualizations, and best practices that help an organization make decisions. Power BI is also one of the leading BI solutions in the market, and many have ranked it as the best BI solution out there. Although the ranking is quite subjective, still, it is fair to say that Power BI is considered as a mainstream solution in the BI domain.
Architecture: Let’s talk about the high-level architecture of Power BI. To demonstrate this, here is a diagram showing the various components of Power BI. If you’re already familiar with Power BI, you should notice that I have excluded Power BI Report Server from the diagram. While it is also a part of this eco-system, the only major difference between Power BI Report Server and Power BI Service is that the former is on premise report sharing platform, whereas the latter is cloud-based report sharing platform. With that said, here is the diagram. Let’s break down this architecture diagram. Usually, these are the components of a report in Power BI:
- Data Source – Power BI connects to a variety of data sources and uses the data from them to create reports. It can connect to databases (SQL Server, Redshift, Oracle SQL Server etc.), spreadsheets, JSON, XML, Sharepoint Folder, and many more. If you want to read more about the compatible data sources, click the link here.
- Power BI Desktop – This is the desktop application that is used by developers to ingest the data, process the data (data transformations and modelling), create visuals and then publish the report to the cloud (Power BI Service), or on-premises server (Power BI Report Server). Power BI Desktop is primarily used for developing the report, so you would find all the options here that could be used to create a report specific to your requirement. This application is only available on Windows, so if you are using MacOS, you might need to install a VM and run the app there.
- Power BI Service – This is the cloud platform that allows you to share reports with the stakeholders. The developers create reports using Power BI Desktop, and then publish them to Power BI Service so that the end user can generate insights, in order for them to make data-driven decisions. Power BI Service has various features, that allow you to create workspaces, configure dashboards, workspace apps, dataflows, and much more. Also keep in mind that Power BI Service is not for developing BI reports, so if you have any major changes that you need to make in the existing report, it will be done on Power BI Desktop.
- Browser & Mobile Apps – Once the report has been published to Power BI Service, users can easily view it using their web browser, or a dedicated app that is present on Android or IOS devices. To see the reports, users would need to have access to their accounts.
Creating and Sharing a Report in Power BI:
To create a new visualizations report, you would need to use the Power BI Desktop app, because as we mentioned earlier, this app is primarily used for report development purposes. If you don’t have the app installed, you can easily download it from Microsoft’s website, or you can download it from the Microsoft Store. Personally, I prefer the store option since it allows for automatic updates on the app. Once you open the app, you are welcomed with a beautiful UI of Power BI Desktop, which looks like this. As you can see, we have the blank canvas at the center where you place all of your visuals, on top we have various options where you connect to the data sources, go to the Power Query Editor (which we shall talk about later), add Measures/Calculated Columns, go to view tab, etc. On the right, you have standard visual types which you can drag to the canvas to create a visualization. Apart from the standard visuals, you also have the option to download custom visuals from the built-in store. These custom-visuals are made by developers from around the world and are either paid, or free. On the left, you have three different views which are report view , table view , and modeling view . The report view is primarily used for creating visualizations, the table view for looking at the loaded data, adding calculated columns or tables, changing data types, and modeling view for creating relations between tables, hiding certain fields/tables, etc.
Now let’s jump to the first key component of creating a report, which is connecting to a data source. In Power BI Desktop, you can connect to a variety of data sources and create report using them. For reference, here is a snapshot of some of the data sources that you can connect to.
As you can see, the users can connect to Excel, XML, JSON, SQL Server, Oracle database, Azure Data Sources, and much more. You can also search for the data source you are looking for, since this window scrolls down to a lot of options. This goes to show that Power BI Desktop is compatible with majority of the data sources present out there.
Once the data source is connected, you can either start developing the report, or you can transform the data by using a built-in tool called Power Query Editor. This tool is built-in to the Power BI Desktop App and is one of its most important parts, since it allows you to clean and transform the data. Power Query Editor performs all the data processing using a language called M-query. Here is a brief overview of the UI of Power Query Editor. We won’t be going into the details of the tool, but this tool offers a lot of features that can be useful to your requirements. Power Query Editor performs standard transformations, like changing data types, performing arithmetic/string/date operations, joins, group by, handling missing values, and much more. You can also implement Machine Learning and AI techniques on your data to generate various insights. On top of all of this, you can even write Python or R scripts on your data set to handle various issues that might not easily be done with the standard options available on Power Query Editor. All of these options can be used with the help of few clicks (Of course, Python and R would require script writing), and Power Query Editor automatically translates those transformations into equivalent M-Query Code. It also allows you to edit the M-Query, but usually that is done by more advanced users who are comfortable with the tool.
As you finish the data processing in Power Query Editor, you can load your data back to Power BI Desktop, where you model the data, and ultimately, create the report which you can share with the stakeholders. Once you are inside Power BI Desktop, you can model the data, by which we mean that we create relations between tables using common fields, and create Measures (which are functions that return scalar values), Calculated Columns, and Calculated Tables by using a language called DAX (Data Analysis Expressions). DAX language is used to perform various calculations within the report and can become quite complex based on the requirement you’re trying to fulfill. Unlike M-Query that we talked about earlier, DAX requires a lot of time and patience to become good at, as it is considered one of the hardest languages to master, since there are a lot of functionalities that it provides and based on the scenario, you are learning something new every time. But you should not worry too much about ‘mastering’ DAX, because naturally you become good at it over time and can navigate through problem relatively easily (Googling problems helps quite a lot though).
After modeling the data, you can start creating visuals inside the report canvas. You can use features such as bookmarks, tooltips, drill-down, drill-through in your report to make the experience more interactive. Just like QuickSight, you can also add multiple pages/sheets where you can group together a bunch of visuals that represent a certain analysis. You can format the visuals based on specification, use built-in machine learning and AI techniques, as well as create visuals with the help of R and Python. Once the report is created and ready to be shared, you can publish it to Power BI Service where you can collaborate with your colleagues (such as Quality Assurance Engineers, other developers) to finalize the report, and share with the actual consumers – the stakeholders.
To summarize the experience of Power BI, it is definitely a bit complex compared to AWS QuickSight and takes some time to getting used to. Overall, it is a great solution if you want to build detailed reports.
Pros and Cons: Now that we have highlighted some key features of Power BI, let’s talk about the advantages and disadvantages of using the software. Let’s first look at some of the pros.
- Availability / Affordability – Power BI in general is quite affordable to use. The desktop app is free for everyone, so if you want to learn about the tool and work on projects, you can easily do so without paying anything. However, if you want to use Power BI Service and all its features, you would need to purchase a pro license at minimum, starting at $13.70. For more details on pricing, please visit the mentioned link.
- Abundant Data Sources – Power BI connects to a variety of data sources, and has great integration with Microsoft’s proprietary services such as Excel, SQL Server. If you ever come across a data source that you’ve not heard of, and want to see if it is compatible with Power BI, there is a good chance it will be available in the list.
- Monthly Updates – The good thing about Power BI is that it is updated every single month. The developers over at Microsoft are constantly adding or improving features all the time, so over time Power BI has become a refined product.
- Mainstream BI Tool – Since Power BI is one of the leading BI solutions out in the market, the support in the community is quite impressive. If you ever run in to a problem, there is a good chance other have come across it too, and you can easily google the problems to find their solutions. On top of this, there are tons of resources online from where you can learn about the tool. You can watch videos on YouTube, read articles, study courses on various websites to learn more about the tool. Personally, I use SQLBI and DAX Guide quite a lot, and also YouTube Channels such as Curbal to learn more about the features within Power BI. This tool has also become a major requirement in the market if you want to become a Data Analyst / BI Developer. So, if you know how to work on the tool, it would definitely help you stand-out in the interview process.
- Custom Visuals – One key advantage of using Power BI Desktop is its ability to use Custom visuals. If your requirement cannot be fulfilled with the default visuals, you can always visit the store and search for the visual type you’re looking for. Although I would mention here that it takes time to search for a particular visual, it’s still safe to say that the tool has a variety of visuals to choose from.
Here are some of the cons of using Power BI. Of course, there are minute details, but these are the major issues I can think of:
- Takes Time to Learn – As Power BI comes with a lot of features, it definitely takes some time to become good at it. You are constantly learning new things as you encounter various situations. Power BI is not just a simple drag-and-drop type of BI tool; it comes with a complete suite of products and features. So, learning most of the things comes with experience and effort.
- Performance – One thing I have noticed over time with Power BI is the performance. With big data, your reports can become quite slow. In order to reduce the issue, you have to optimize the reports by applying various techniques. Applying the optimization techniques requires knowledge and experience, so if you’re new to the BI domain, optimization can become a major hurdle and you can end up with a report that the end-user can’t even see.
Power BI vs QuickSight: The Comparison
Now that we have highlighted the key features, as well as pros and cons of using the two BI solutions, we are ready to make a brief comparison between the two. To make things simple, I have created this table which highlights the major differences between them. The two BI solutions that we have discussed here have their unique features and target markets. Both have their ups and downs, and if I have to pick a tool, I would definitely choose Power BI because of the reasons I have highlighted in the article. Power BI definitely fits the need of majority of the users, but you can always use QuickSight if your reporting is simple and you do not require all the features that Power BI provides. We do, however, hope that QuickSight covers up for the lost time and catches up to its competitors by adding new features consistently over time, so that it challenges the top contenders and takes a fair share of the market.