Time needed to read 11 minutes
Us and Gartner Square 2019 BI 🙂
The purpose of this article is to compare the top three BI platforms that are in the top of the Gartner Quadrant:
- Power BI (Microsoft)
— tableau
—Qlik
Figure 1. Gartner BI Magic Quadrant 2019
My name is Andrey Zhdanov, I am the head of the analytics department at Analytics Group (). We build visual reports on marketing, sales, finance, logistics, in other words, we are engaged in business analytics and data visualization.
My colleagues and I have been working with various BI platforms for several years. We have a very good project experience, which allows us to compare platforms from the point of view of developers, analysts, business users and BI system implementers.
We will have a separate article comparing prices and visual design of these BI systems, so here we will try to evaluate these systems from the point of view of an analyst and developer.
We single out several areas for analysis and evaluate them according to a 3-point system:
— Entry threshold and requirements for the analyst;
— Data sources;
— Data cleaning, ETL (Extract, Transform, Load)
— Visualizations and development
— Corporate environment — server, reports
- Support for mobile devices
— Embedded (built-in) analytics in third-party applications/sites
1. Entry threshold and requirements for an analyst
Power BI
I've seen a lot of Power BI users who weren't IT professionals but could create a decent report. Power BI uses the same query language as Excel - Power Query and the DAX formula language. Many analysts know Excel well, so it is quite easy for them to switch to this BI system.
Most of the actions are fairly easy to perform in the Query Editor. Plus there is an advanced editor with the M language for professionals.
Figure 2 - Power BI Query Builder
QlikSense
Qlik Sense looks very friendly - a small number of settings, a quick ability to create a report, you can use the data load constructor.
At first, it seems easier than Power BI and Tableau. But from experience I will say that after a while, when an analyst creates a couple of simple reports and needs something more complex, he will face the need to program.
Qlik has a very powerful data loading and processing language. It has its own Set Analysis formula language. Therefore, an analyst must be able to write queries and connections, place data in virtual tables, and actively use variables. The possibilities of the language are very wide, but it will require learning it. Probably all Qlik analysts that I know have some kind of serious IT background.
Qlik integrators, like us, often like to talk about the associative model, when when data is loaded, all of it is placed in RAM, and the connection between the data is carried out by the platform's internal mechanism. That when choosing values, no internal subqueries are made, as in classic databases. Data is issued almost instantly due to pre-indexed values and relationships.
True, in practice this leads to the creation of automatic table joins, if the field names match. For example, you cannot have different tables without relationships that will have the same field. You have to get used to it. You either have to rename the columns and make sure that the names do not match, or reduce all the fact tables into one and surround them with star references. It’s probably convenient for beginners, but for experienced analysts it doesn’t matter.
A typical interface for loading and processing data for an analyst looks like this.
Figure 3. Qlik Sense data load editor, Calendar table
Note: In Power BI, the situation usually looks different, you leave different fact and lookup tables, you can manually join tables in the classic way, i.e. I map columns to each other manually.
Artwork
Developers position Tableau as BI with a convenient and friendly interface that will allow the analyst to independently study their data. Yes, we had analysts in our company who, without IT experience, could make their own reports. But I'm lowering Tableau's rating for several reasons:
- Weak localization with Russian language
— Tableau Online servers are not located in Russia
- A fairly simple load constructor starts to cause problems when you need to build a rather complex data model.
Figure 4. Tableau Data Load Builder
One of the interview questions we ask Tableau analysts is “How to build a fact table model with lookup tables without bringing everything into a single table?!” Data Blending requires thoughtful use. I've corrected my analysts many times for duplication of data after such joins.
Plus, Tableau has a rather peculiar system when you make each chart on a separate Sheet, and then create a Dashboard, where you start placing the created sheets. Then you can create a Story, this is a combination of different Dashboards. Development in Qlik and Power BI is easier in this regard, you immediately put chart templates on the sheet, set measures and dimensions, and the Dashboard is ready. It seems to me that the labor costs for preparing in Tableau are increasing because of this.
2. Data sources and loading
There is no clear leader in this section, but we will highlight Qlik because of a couple of nice features.
Tableau in the free version is limited in sources, but we focus more on business in our articles, and businesses can afford commercial products and analysts. Therefore, Tableau did not reduce the rating for this parameter.
Figure 5. List of possible Tableau sources
Otherwise, the list of sources is impressive everywhere - all tabular files, all standard databases, web connections, everything works everywhere. I have not come across non-standard data stores, there may be some nuances there, but in most cases you will not have problems loading data. Exception, perhaps only 1C. There are no direct connectors to 1C.
Qlik partners in Russia sell their own connectors for 100 - 000 rubles, but in most cases it is cheaper to upload from 200C to FTP to Excel or a SQL database. Or you can publish the 000C database on the web and connect to it using the Odata protocol.
PowerBI and Tableau can do this as standard, but Qlik will ask for a paid connector, so it’s also easier to upload to an intermediate database here. In any case, all questions on connections can be resolved.
Figure 6. List of possible Qlik Sense sources
Additionally, it is worth noting the feature of Qlik that they supply both paid and free connectors as a separate product.
Figure 7. Additional Qlik Sense connectors
From experience, I will add that with large amounts of data or numerous sources, it is not always advisable to connect a BI system right away. In serious projects, a data warehouse, a database with data already prepared for analysis, etc. is usually used. It is impossible to take and upload, say, 1 billion records to the BI system. Here it is already necessary to think over the architecture of the solution.
Figure 8. Power BI data sources
But why was Qlik spun off? I really like 3 things:
- QVD files
Own data storage format. Sometimes it is possible to build serious commercial projects only on QVD files. For example, the first level is raw data. The second level is processed files. The third level is aggregated data, etc. These files can be used in different applications, different employees and services can be responsible for them. The download speed from such files is ten times faster than from conventional data sources. This allows you to save on database costs and share information between different Qlik applications.
— Incremental data loading
Yes, Power BI and Tableau can also do this. But Power BI will require an expensive Premium version, and Tableau doesn't have the flexibility of Qlik. In Qlik on QVD files, you can take snapshots of systems at different times and then process this data as you like
- Connecting external scripts
In addition to QVD files for storing data, in Qlik the script code can also be moved outside the application and connected with the Include command. This already allows you to organize teamwork, use version control systems, manage a single code for different applications. Power BI has an advanced query editor, but we couldn’t set up such teamwork as in Qlik. In general, all BI have problems with this, it is simply impossible to manage data, code, and visualizations in all applications from a single place at the same time. The maximum that we managed to do was to take out QVD files and script code. Visual elements have to be edited inside the reports themselves, which does not allow us to massively change visualizations for all clients at the same time.
But what about such a mechanism as Live connection? Tableau and Power BI support LIVE connections to a number of sources, unlike Qlik. We are rather indifferent to this feature, because practice shows that when it comes to big data, it becomes simply impossible to work with a LIVE connection. And in most cases, BI is needed for big data.
3. Data cleaning, ETL (Extract, Transform, Load)
I have 2 leaders in this section, Qlik Sense and Power Bi.
Let's just say Qlik is powerful but complicated. Understand their SQL-like language and you can do almost everything - virtual tables, table joins and joins, loop through the table and generate new tables, a bunch of string processing commands. For example, a field in 1 cell, which is filled with data like “Ivanov 851 Bely” can be decomposed on the fly not only into 3 columns (as everyone knows how), but immediately into 3 lines, for example. It is also easy to do the same thing on the fly by combining 3 lines into 1.
Figure 9. How to load and transpose a table in Qlik Sense from Google Sheets
Power BI is simpler in this regard, but most tasks are easily solved through the query designer. I set a number of parameters, transposed the table, worked on the data, and all this without a single line of code.
Figure 10. How to load and transpose a table in Power BI from AmoCRM
Tableau seems to me to have a different ideology. They are more for beauty and design. It seems to me very difficult to connect a bunch of different sources, combine them all and process them inside Tableau. In commercial projects, in most cases, Tableau already prepares and accumulates data in storages and databases.
Figure 11. How to load and transpose a table in Tableau
4. Visualizations
In this section, we did not single out the leader. We will have a separate article, where we will show the same report in all 3 systems using the example of one case . It is more a matter of taste and skill of the analyst. On the Internet, you can find very beautiful pictures built on the basis of any of these systems. The basic visualization capabilities are about the same for everyone. The rest is solved by using Extensons. There are paid ones and free ones. There are extensions from the vendors themselves, as well as from freelancers and integrators. You can write your own visualization extension for any platform.
I like the Tableau style, I think it is strict and corporate. But getting a really beautiful picture in Tableau is hard. A great example of Tableau visualization using the same extensions. I will not be able to repeat this, because. I don't have these extensions, but it looks nice.
Figure 12. Appearance of Tableau reports with Extensions
Power BI can also be made interesting.
Figure 13. Appearance of Power Bi reports with Extensions
The only thing I don't understand about Power BI is why they have such weird default colors. On any chart, I am forced to change the color to my corporate, corporate one and am surprised at the standard coloring.
Qlik Sense also depends on Extensions. The use of add-ons can change the reports beyond recognition. You can also roll up your theme and design.
Figure 14. Appearance of Qlik Sense reports with Extensions
From a developer's point of view, Qlik Sense is more convenient for me because of the standard options as alternative dimensions and measures. You can set several dimensions and measures in the visualization settings, and the user can easily set what to look at on a particular chart.
In Power Bi and Tableau, I have to configure parameters, buttons, program the system behavior depending on these parameters. I wonder why it's so difficult. The same goes for the ability to change the appearance of viloisilizations.
In Qlik, you can hide different types of visualizations in one object, but in Power BI and Tableau this is more difficult. Again, it depends more on the skill of the performer. You can make a masterpiece in any system, but without experience, inexpressive graphics will be obtained everywhere.
5. Corporate environment - server, reports
All products have corporate server versions. I have worked with all editions and I can say that all have strengths and weaknesses. The choice of product should be selected from your requirements for software, taking into account their nuances. All vendors can distribute rights both at the level of accounts and groups, and at the level of Data Row Level Security. Scheduled automatic refresh of reports is available.
Qlik Sense Enterprise is a great opportunity to build analytics inside your organization for medium businesses. It may seem more expensive than Power BI Pro, but do not forget that Power BI Pro servers are in the cloud on Microsoft territory and you cannot influence performance, and when you need Power BI Premium, which can be deployed on your servers, then the price starts from $5000 per month.
Qlik Sense Enterprise starts from 230 rubles. for 000 licenses (fee per year, then only technical support), which is much more affordable than Power BI Premium. And Qlik Sense Enterprise will allow you to use all the features of Qlik. Perhaps except for one. For some reason, Qlik decided that a feature such as the ability to send PDF reports via email should be provided as a separate NPrinting service.
But Qlik Sense Enterprise is more powerful than Power BI Pro and therefore the following comparison can be made here.
Qlik Sense Enterprise = Power BI Premium, with equal features, it turns out cheaper with medium implementations. Large implementations are usually calculated already on the vendor's side, where they can provide individual conditions for your company.
In this regard, we will give preference to Qlik Sense Enterprise, there are all the possibilities to build serious analytics on huge data. In our opinion, Qlik will work faster than Power BI on large arrays, at Qlik conferences there were clients who first tested their data in billions of records and Power BI showed worse results.
Figure 15. Appearance of Qlik Sense Enterprise server reports
Qlik Sense Cloud = Power BI Pro. Qlik Sense Cloud is 1.5 times more expensive* and there is a very significant limitation that this platform does not allow us. You can't use Extensions, even built-in ones. And without extensions, Qlik loses some of its visual beauty.
Figure 16. Appearance of the Power BI Pro dashboard
*An alternative is to use Qlik Sense Enterprise on a subscription basis. But so that this article is not perceived as advertising, we will not cover our pricing.
And Tableau stands a little aloof for us. They have both cloud subscriptions of $70 for a developer and $15 for a viewer, as well as expensive server solutions. But the main idea of Tableau is that for big data it is necessary to organize data processing and storage on the side. Objectively less functionality does not allow for serious data processing in Tableau. Visualize, analyze, yes. But for small and medium businesses, creating a separate storage is usually problematic. I'd downvote Tableau so if it wasn't for their 1 chip. The Tableau server sends out scheduled emails with CSV or PDF attachments without any problem. Moreover, you can distribute rights, autofilters, and so on. For some reason, Power BI and Qlik cannot do this, but for some it can be critical. Due to this, Tableau maintains a position in our dispute.
Figure 17 Tableau Server Control Panel Appearance
Also in a corporate environment, you need to think about the cost of implementation and maintenance. In Russia, there is a practice that Power BI is more common in small businesses. This led to the emergence of a large number of vacancies and resumes, the emergence of small integrators. This will allow you to find specialists for a small project. But most likely, all of them will not have experience in large implementations and working with big data. Qlik and Tableau are the opposite. There are few Qlik partners, and even fewer Tableau partners. These partners specialize in large implementations with a large average bill. There are not many vacancies and resumes on the market, the threshold for entering these products is more difficult than in Power BI. But in Russia there are successful implementations of these products for thousands of users, and these products perform well on big data. You just need to understand the strengths and weaknesses of the products in relation to your business.
6. Support for mobile devices.
In this section, we will highlight Power BI and Tableau. You can install mobile applications and they will look quite adequate on the screens of mobile devices. Although it seems to us that analytics on mobile devices is losing to analytics on a PC. Still, it is not so convenient to use filters, the pictures are small, the numbers are hard to see, etc.
Figure 18. Power BI report appearance on iPhone
Figure 19. Tableau report appearance on iPhone
Figure 20. Qlik Sense report appearance on iPhone
Why have Qlik scores been lowered? For reasons unknown to us, the mobile client is only available on the iPhone; on Android, you will have to use a regular browser. Plus, when using Qlik, you immediately have to understand that a number of Extensions or visualizations are not reduced or auto-positioned in mobile devices as expected. A report that looks very nice on a PC looks much worse on a small screen. You have to make a separate report for mobile devices, where to remove filters, KPIs and a number of other objects. This also applies to Power BI or Tableau, but in Qlik it is especially pronounced. We hope Qlik will work on their mobile client in the future.
If you plan to spend a lot of time doing analytics from mobile devices, then it makes sense to install all 3 clients and check their display on test reports. Any vendor on their website has a gallery of test reports for review.
7. Embedded (built-in) analytics in third-party applications / sites
Using analytics as a third-party service is not always convenient. Perhaps you are developing your product, but are not ready to start developing visualization and analytics engine from scratch. Perhaps you want to deploy analytics on your website so that the client registers himself, uploads his data and conducts analysis inside his personal account. To do this, you need embedded analytics (Embedded).
All products allow you to do this, but in this category we will single out Qlik.
Power Bi and Tableau clearly say that for such purposes you need to buy a separate product, Tableau Embedded Analytics or Power BI Embedded. These are not cheap solutions in thousands of dollars per month, which immediately limits their use. Most projects immediately become unprofitable for our clients. This means that you need to not only publish a report on the entire Internet, but ensure that reports are published according to certain accesses, with data protection, user authorization, etc.
And Qlik will let you get out. Of course, they also have Qlik Analytics Platform which is licensed per server and organizes an unlimited number of connections. It will also be expensive like the competitors of Tableau and Power Bi. And in the case of unlimited connections, there are not many options.
But in Qlik there is such a thing as Mashup. Let's say you have Qlik Sense Enterprise and 10 licenses. Standard analytics, appearance, everything is already tired. You raise your own website or application, and you can implement all your analytics right there. The trick is that, simply speaking, Mashup is a visualization in the program code. Using the API, you can programmatically create a visualization within your application or site. You will still need Qlik Sense Enterprise for licensing (licenses to connect to the site = licenses to connect to BI), to load data, etc., but the visualizations will no longer be displayed on the side of this server, but embedded in your application or site. You can use CSS styles, set new fonts, colors. Your 10 users will no longer log into the analytics server, but will use your corporate portal or application. Analytics will reach a new level.
Figure 21. Appearance of a Qlik Sense report embedded on a website
It will be difficult to understand where the elements of the site are, and where Qlik Sense already begins.
Of course, you will need a programmer, and even more likely several. One for web programming, one for working with the Qlik API. But the result is worth it.
Conclusions. Let's summarize.
It is difficult to say unambiguously who is better and who is worse. Power BI and Qlik are on par in our competition, Tableau is slightly inferior. But the result may be different for your business. In BI platforms, the visual component is very important. If you have looked at dozens of demo reports and pictures on the Internet for all BI systems and you do not like how one of the platforms looks, then most likely you will not implement it, even if you are satisfied with the price or technical support. characteristics.
Next, you will definitely need to calculate the cost of licenses, implementation and maintenance of the BI platform. Perhaps in your case the leader will be revealed. Of great importance is the contractor or the ability to hire the right specialist. Without professionals in any platform, the result will be disastrous.
Good luck with your BI integrations, Andrey Zhdanov and Vladimir Lazarev, Analytics Group
Source: habr.com