I wrote an article about how to choose BI before ""BI selection secret" BI tools don't know how to choose? You must collect this model selection score sheet! ".
After the article was published, many friends sent me private messages, wanting me to publish an evaluation of domestic BI tools. The first issue will start today: Fanruan VS Guanyuan
Still put the evaluation elements first:
(1) Tools
① Ease of use
Most of the targets of BI are business personnel and professional data analysts. Therefore, whether the tools are easy to use, whether the interactive experience is smooth, and whether there are enough learning resources largely determine whether users will Continuous use, and whether it can be used in depth. Most BI tools on the market are based on the concept of "zero code": they can be dragged, pulled, and implemented without writing complex functions. Fanruan's Fine BI and Guanyuan's BI are no exception.
The functions of Guanyuan are presented through layers of pages, so the structure of each page is very simple. Even novices who have never been exposed to BI tools can follow the prompts for data analysis and chart production.
However, the Fine BI operation interface has a relatively large amount of information, and the functions are tiled on a single page, which is distinguished by lines and surfaces. It is difficult to quickly analyze data when the screen is small and the amount of data is large. But it is also because the display of functional information is relatively rich and there are relatively few levels. When analyzing real scenes, you will feel that the operation efficiency is particularly high.
for example:
I want to analyze the sales data of store A this month, so my first step is to find these data.
In Guanyuan BI, I only need to click three times, and the page will find the store data table layer by layer. This is the embodiment of the ease of use of Guanyuan BI. But pay attention to the premise of this operation: I know the full name of the table, and I also know which fields are in the table.
In Fine BI, I search directly or enter keywords in the search box to find the corresponding table and data. When I find that this table is not what I want, I can directly select another one to view until I find the data I want. This is The Fine BI operation interface is rich in information.
In the actual situation, there are more scenarios like the following. I only know the general keywords, but I don’t know which table it is. Therefore, I need to open several tables one by one to determine where the data I want to use is. inside a table.
In terms of whether it is easy to use and interactive experience, you will feel that Guanyuan is better when you watch the product demonstration, but in the actual business use process, you will find that the interaction of Fanruan's Fine BI is more suitable for exploratory analysis.
In fact, for most people, the habit of data analysis comes from Excel. Fine BI continues Fanruan’s reporting tool FineReport, and uses an Excel-like interface, while Guanyuan’s smartETL operator does not have many operators. Simple, a lot of operations may be required in the process of actually processing data, and it is still difficult to use it in practice.
Fanruan and Guanyuan both have text version and video version tutorial entrances on their official websites, and the learning materials are the same.
②Performance
In terms of product architecture, Guanyuan packages Clickhouse, and Fanruan uses StarRocks big data solution. In the case of "high concurrency", a query in clickhouse occupies half of the CPU of the server, and starpocks can perform pre-aggregation. Hundreds of millions of data tables can be aggregated into several million tables first to improve concurrency efficiency. Generally speaking, ClickHouse is suitable for scenarios where there are few dimension changes and wide tables. StarRocks not only performs better in single-table tests, but also has greater advantages in multi-table association scenarios.
For specific selection, please refer to this article "ClickHouse vs StarRocks Selection Comparison".
When it comes to data processing, it is best to divide the performance into two parts: below the 10 million level and above the 10 million level (large data volume generally refers to the 100 million level or above).
For datasets with less than 10 million rows, the Spider engine used by Fanruan and the Spark engine used by Guanyuan can provide a better response experience.
But in the scenario of large data volume, don't listen to the bragging of the manufacturer. The domestic and foreign BI manufacturers I know can't provide better support, and it is no longer suitable to use the extraction mode for analysis. In this case, either optimize the underlying data and build a big data platform; or find out whether the manufacturer has a corresponding solution.
As I understand it, Fanruan has a large data volume solution, which probably picks a database with good performance for adaptation and optimization, so as to support billion-level big data. How is the specific implementation effect? I haven’t learned about it here. Friends who have learned about it can share it in the comment area.
③Functional
BI tools have many functions, and different tools have different focuses. Therefore, on the basis of ensuring ease of use and performance, it depends on the matching degree of functions and enterprise needs.
data preparation
Both Fanruan and Guanyuan support the two data connection forms of filling in and connecting to the database. Guanyuan directly has the function of data reporting in BI, while Fanruan data reporting requires the use of the reporting tool FineReport, which has a strong reporting ability, but product integration needs to be investigated.
Besides, when it comes to connecting to the database, the access difficulty of both is relatively simple, but there are some differences in the types of databases supported.
In terms of access, as long as the database server adds a whitelist that allows access, fill in the corresponding database address, account password, and then you can access immediately. Fine BI supports most mainstream databases, more in number than Guanyuan.
The steps of connecting Guanyuan and Fanruan are the same, and both are relatively simple. Guanyuan basically supports common databases on the market, and if there are no special needs, all of them are enough.
data processing
In the ETL link, Fanruan uses the family bucket model as always, and has a special tool FineDataLink to provide a one-stop data processing platform. It can be said that it has prepared a data development and processing module for IT, supporting data synchronization, data conversion, comparison, deletion, and data processing. Professional data processing operations such as association, row and column conversion, json parsing, SQL script, conditional branch, loop container, etc. In essence, Fanruan believes that data processing should be handed over to IT, while data editing and analysis can be handed over to the business itself.
Guanyuan has a dedicated intelligent data processing module (Smart ETL), and users can drag and drop to perform data processing operations and data fusion on multiple data sources. It can be seen that Guanyuan wants to provide simple ETL tools to let the business undertake part of the data processing work. However, in the actual use process, more formulas are used for text processing and index calculation, and the quality requirements for business personnel are relatively high.
Data Analysis and Visualization
I won’t expand on the basic data analysis functions. There is not much difference between the two tools, such as OLAP analysis. Both can easily configure linkage, jump, drill, etc.; both support parameters to participate in calculation and filtering; both support fast calculation Such as one-click calculation of year-on-year ratio, maximum and minimum, average, variance standard deviation (not supported by Guanyuan), proportion, accumulation, etc.
At the level of advanced data analysis, FanRuan Fine BI6.0 version strongly launched the DEF function for complex analysis scenarios, claiming that mastering one formula can solve all data analysis scenarios. The emergence of the DEF function can be regarded as the first shot to impact advanced computing in China. It is somewhat declaring war on Tableau LOD and DAX CALCULATE. I will study more and try to add some practical applications next time.
In terms of visualization, Fanruan Fine BI has the same design ideas as Power BI and Tableau. They both design visualization modules based on graph grammar, and decompose the graph construction process into a series of composable elements and operations, so as to achieve repeatability, scalability and Flexible graph generation.
In addition, Fine BI can create a visual dashboard, but it does not have the function of large-screen visualization (usually implemented with FineReport), nor does it have the function of complex reports (usually implemented with FineReport). Therefore, under normal circumstances, the company purchases Fanruan products in the form of FR+BI.
In terms of visualization, Guanyuan BI chooses the idea of "graphics before problems", first determines the type of charts needed, and then displays them. It is slightly weak in the flexibility of style switching. For this reason, they also have a fixed number of charts. The input has more than 50 fixed chart types.
In addition, Guanyuan's BI fully integrates the large-scale visualization and reporting functions, and only divides into functional modules, without distinguishing between software, and it is better in terms of integration.
Platform control
Authority management and control is a point of concern in normal procurement. I have sorted out the functions of relevant authority management and control here, and you can take a closer look when you compare them. Here are two issues I have noticed. Fanruan's permissions do not support template reuse directly, but reuse permissions through the linkage between tables, which is somewhat different from our usual usage habits. However, Guanyuan’s row authority control needs to use sql. The more layers there are, the more complex the sql will be. It will be uncomfortable when the company structure is more complicated.
Let’s take a seemingly extreme and common example: A’s leader manages the authority of a certain data table uploaded by C of Department A. In a single project, A and B come from two departments, but C is not the leader of these two departments. , At this time, it is necessary to set the permission allocation, and writing sql can write to crash.
(2) Vendor
Fanruan started early and is well-known, with a total of more than 26,000 customers. It used to have a place in the report field with FineReport, and now it is promoting Fine BI6.0 more. It is a relatively old-fashioned software company that focuses on "never going public", and hardly has to worry about being cut off. The manufacturing, financial industry, and pharmaceutical industry have more solutions and customers, and the volume is very loud.
Guanyuan can be regarded as a rising star in the BI industry, taking the route of financing and listing, focusing on the retail industry. In recent years, the market expansion of Guanyuan has been relatively intense. The solution side is full of senior businesses that are deeply involved in the industry. The marketing side has also proposed the concept of "AI+BI", and there is little movement on the product side.
From my point of view, when selecting a company, it is also necessary to consider whether it can be implemented. I was invited to participate in the "AI+BI" internal test activity before, and the error rate is still relatively high. However, I am optimistic about "AI+BI" and hope that domestic BI can generate international buzz.
(3) Market
In Gartner's "2021 Magic Quadrant for Analytics and BI Platforms" report, both Power BL and Tableau are in the leader quadrant. The leader in the domestic market is Fanruan Software, which has topped the IDC China BI Market Tracking Report for many years.
summary
Under the wave of localization, domestic manufacturers have obvious ambitions to replace foreign software, and indeed they have more advantages in localization services and landing capabilities. It can be foreseen that, under the gestation of a large number of favorable policies and innovative soil, more domestic BI products will appear in the domestic market to compete with foreign products, while domestic BI products will receive more attention.
Finally, I would like to share with you a BI tool selection scoring table. The company compares the above factors according to the free situation, scores according to the evaluation dimensions, and then calculates the total score according to the weight.
Reprinted from miaojun