"After using DataEase, we pointed out some inconveniences we found during the use process. The project team and users worked together to find solutions and optimizations to make this product better. DataEase is an open source product jointly created by developers and users."

——Mr. Ma, a core user of the DataEase open source community

Editor's Note: The following content is based on the community sharing of Mr. Ma, a core user of the DataEase open source community.

Mr. Ma works for a global social entertainment company. The daily business operation of the company involves a large number of data information such as daily activity and monthly activity on the platform. From 2021 to the present, the company has selected three BI tools according to its own business operation needs, namely Quick BI, Superset and DataEase, and the selection and elimination of each product are closely related to the company's needs.

■ Quick BI period : Before August 2021, the company has not established a complete data warehouse, and there is only one person in the company responsible for data visualization, and the product used is Quick BI Personal Edition. At that time, most of the data was sent to the business department in the form of emails, but the email data was static data, and historical data could not be directly queried. Historical data required querying historical emails, which was labor-intensive;

■ Superset period : With the development of the company's business, the company gradually established a mature data department, built its own data warehouse system, and began to use Superset. Compared with Quick BI, the biggest advantage of Superset is that it is an open source software, which does not need to be shared with business departments by email after downloading data or taking screenshots of pages. Just log in to the same system, and the data completed by the data department can be presented to the business department, and the business department can check the historical retained data by itself. Whether it is detailed data or data of a single dimension, you can view historical data. On this basis, they completed the first leap from the Quick BI period (email requirements) to the Superset period—that is, the transition from static data to dynamic data;

■ DataEase period : After completing the first transformation, Mr. Ma's data team began to think about the second problem. Although Superset is an open source software, it is a foreign open source project after all, which does not conform to the usage habits of Chinese people, and cannot fully meet the needs of the company in terms of functions. For example, the technical department needs to create a large display screen for performance monitoring data, which is very troublesome to implement through Superset. So the data team began to look for the next BI product that is more suitable for their company. At this time, they stumbled upon DataEase, a domestically produced open source visualization product. This product perfectly solved two problems encountered by Mr. Ma's team when using Superset: it did not meet the usage habits of Chinese people, and it was difficult to handle real-time data refresh. So, they made the second leap—from a usable product to an easy-to-use product.

During Mr. Ma's BI product upgrade process from Quick BI, Superset to DataEase, the DataEase open source community also witnessed his journey from product user to product contributor as he gradually joined the open source world.

Community User Interviews

What kind of opportunity made you start contacting and using DataEase?

Mr. Ma: Our company has been using Superset as a data chart display system before. With the growth of business, many operations and product colleagues want to configure reports by themselves. After they contacted Superset, they generally reported that Superset configuration is cumbersome. For example, when configuring a line chart, Superset needs to configure time columns, time granularity, indicators, dimensions, grouping, etc. Many business colleagues are confused in the step of configuring the chart, and it is difficult to understand the terms dimension and grouping. In fact, what business colleagues need is an easily configurable chart system that can integrate data sets, charts, dashboards, and configuration items. This kind of chart system has a low learning cost and allows business colleagues to get started quickly.

After a series of product research, DataEase was discovered. I tried it on the test environment on the DataEase official website, and found that it was quite easy to use, so I started to deploy it to the business department.

▲ Figure 1 Superset view design page

▲ Figure 2 DataEase view design page

■ Introduction to DataEase View Design Ribbon

[Number 1]: switch between data operation area, style editing area and advanced function area

[Number 2]: Search

[Number 3]: Field editing

[No. 4]: Replace the data set

[No. 5]: Optional dimension list

[No. 6]: List of optional indicators

[No. 7]: Chart type (including ECharts and AntV)

[No. 8]: Result display

[Number 9]: Dimension setting area

[Number 10]: Indicator setting area

[Number 11]: Result filter

[No. 12]: Drilling dimension setting area

[No. 13]: Chart display

What do you think are the advantages and disadvantages of DataEase?

Mr. Ma: Our company is currently using both Superset and DataEase. The product department mainly uses Superset, and the operation department mainly uses DataEase. In addition, DataEase is mainly used to display UID granularity data.

Compared with DataEase, the advantage of Superset is mainly reflected in its caching mechanism. As more and more reports are created, the problem of increased request volume also arises. Since we use the direct connection method, DataEase's caching effect is not obvious and the response is relatively slow when dealing with hundreds of parallel queries. Superset can customize the expiration time of the cache or not cache. This function point is more practical for us.

The advantages of DataEase are also obvious, including the following points:

1. My first impression of DataEase is that the interface is simple and beautiful . It can easily create various styles of charts. The background pattern and the style border of the chart can be customized. Most importantly, the process of configuring charts is simple enough. For business personnel, he only needs to know what kind of data sets are needed, and he can directly create dashboards. Superset's data set is a messy list structure, while DataEase's data set can be classified hierarchically.

▲ Figure 3 Superset user interface

▲ Figure 4 DataEase user interface

2. The community communication group of DataEase is very active, and the customer service will patiently answer our questions. This is an open source ecological environment, which makes us believe that DataEase is a dynamic project.

3. The community version of DataEase is completely open source , developed with our most familiar Springboot+Vue.js technology stack, which is beneficial for us to do some customized things on the basis of it.

What kind of business needs does DataEase solve for your company? How to use DataEase?

Mr. Ma: In order to make it easier for the business side (such as sales, operations and other personnel who view or make dashboards in the enterprise) to use it more conveniently, we will prepare and classify the data sets in DataEase in advance. The specific operation is that we will create separate folders for each business, and we will continue to create secondary folders according to the region and function under the business. Through such a classification method, the business side can clearly know what data it is.

With the data, the business side can easily configure their own reports, freeing up the manpower of the development team. DataEase also supports sharing reports, users can easily share their own reports with others. Now, if business personnel configure and generate dashboards by themselves, they will choose to use DataEase.

DataEase Community Edition has some limitations, how do you deal with these limitations?

Mr. Ma: Yes, the community version of DataEase does not have the function of open authority control, which is troublesome for companies without secondary development capabilities. Therefore, if you use DataEase as a team, it is best to purchase the Enterprise Edition. The permission control function of DataEase Enterprise Edition can control row permissions and column permissions, which is a very powerful function. In the face of some demand scenarios, for example, a dashboard needs to show different business departments their corresponding data. If it is configured through the row permission function of DataEase, it will be done in a few tens of seconds. At present, we can only make one dashboard for each business department.

Secondly, in response to our own business needs, colleagues in the operation want to monitor the Kanban indicators, so we developed DataEase’s alarm function, which can directly send alarm emails, text messages, phone calls, etc. for indicators that reach the threshold.

In DataEase's template market, users can download templates for free and directly apply them to their own dashboards. Do you think this will bring some convenience to users?

Mr. Ma: I have used DataEase's template market, and it is very convenient. After we download the template, we can directly replace the data to use the template without having to do the styling ourselves. I heard that users can also contribute to the template market, but we have not yet contributed. I have a small suggestion. Most of the templates in the template market are dark-colored templates, but our company uses mostly light-colored dashboards. Because the dark dashboard is visually dazzling, the light dashboard will give people a clean and refreshing feeling.

You mentioned that DataEase's community exchange group is relatively active, so will you and your colleagues also communicate with you or ask questions in the group?

Mr. Ma: I have pulled all my colleagues into the community group. Although we have done secondary development on DataEase, we still don't know many small function points, so colleagues will ask questions in the communication group when they encounter problems. Sometimes I see a question asked by others that I can answer, and I will answer it.

I also fixed several small bugs for DataEase, and submitted two PRs, all of which have been merged into the main branch. Because I have done open source before, I am very clear about the process specification of the entire code development such as PR, running tests, and merging into the main branch. Sometimes we may find some small bugs, but it does not affect the main function of DataEase, so we will submit a PR according to the process.

You now know a lot about the DataEase project. Any suggestions on it?

Mr. Ma: I mentioned before that compared with Superset, one of the shortcomings of DataEase is that the caching effect of DataEase is not obvious. I think this problem can be solved perfectly if we can make a front-end asynchronous. This deficiency mainly occurs in two usage scenarios:

The first scenario is that when the user clicks on a page that has not been loaded, and then clicks on another page, DataEase will not jump immediately, but needs to wait until the previous page is loaded before jumping. For users, it is best to jump directly without waiting for the current page to load;

The second scenario is that DataEase needs to request data no matter when editing the dashboard or dragging dimension indicators, so the operation is not so smooth when the amount of data is large. If a button is set, the dashboard will request data from the background after the user clicks that button, and the overall refresh will be performed after editing, which can reduce requests and make users use it more smoothly.

For cache issues, please refer to other open source BI products. When enterprises choose visualization products, they will have an instinctive distrust, that is, users do not know whether this thing can be used after taking it. In fact, many open source BI products do not have much publicity, but the main reason why they can attract customers is that these products have experienced the practice of many users, so the stickiness of users takes time to settle. Products on the ToB side need to be iterated and updated to attract more companies to use, and then verify the stability of the product. This also confirms the slogan of Feizhiyun: "Software is valuable only when it is used, and only then can there be opportunities for improvement."

The iteration speed of the DataEase project was very fast before. In recent months, because of the 2.0 version, the iteration speed of the 1.0 version will be slower. In fact, I think the current update frequency is quite good. The speed of bug fixes is still very fast, and features can be developed slowly. Because making a product should take into account the majority of users, rather than serving a single user, each requirement should be investigated more comprehensively, and it is necessary to consider whether the new version will affect other users.

Community Sharing｜Enterprise BI Selection: From Quick BI, Superset to DataEase

Community User Interviews

Guess you like