Today , when you look at business intelligence or BI tools in Baidu, you can always see Tableau . It's not that Tableau 's marketing is doing well , but that the domestic cognition and choice of business intelligence tools seem to fall on Tableau . As a result, both inside and outside the industry have biased views on the concept of business intelligence , thinking that it is a front-end display tool and a chart .
There's no denying Tableau here . Indeed, Tableau 's visualization and chart production capabilities are praised by everyone , which is worth learning from domestic BI manufacturers. However , from the perspective of the application of business intelligence , enterprises pay more attention to BI , more on the performance of data processing (data volume, speed, stability), product adaptability (developability, integration) and analysis efficiency (Show effect, operation experience). From this point of view, there is still a lot of room for the development of BI .
So , back to the question, how should domestic BI tools break through?
Let's talk about business intelligence first . The original definition of business intelligence refers to the use of modern data warehouse technology, online analysis and processing technology, data mining and data display technology for data analysis to achieve business value. As a tool, business intelligence is used to process existing data in the enterprise and convert it into knowledge, analysis and conclusions, assisting business or decision makers to make correct and sensible decisions, and helping enterprises make better use of data to improve decision-making quality technologies, ranging from data warehouses to analytical systems. Therefore, business intelligence is strictly speaking a set of solutions , which is based on the existing IT technology structure of the enterprise and provides a solution for fast and accurate data analysis.
From the exploration of FanRuan in the field of enterprise data analysis for more than ten years , combined with the product positioning of Fine BI , we can sniff out the following:
1. Be realistic about the data analysis needs of enterprises
The IT construction of each enterprise is different, the database is diverse , the data is standardized, and the development and integration requirements of the system are also different . In addition to supporting various types of databases and data sources, BI also needs to support big data platforms such as Hadoop , GreenPlumn , and various data warehouses . For some enterprises with data warehouses, while others only have simple databases, some enterprises have a large amount of data, and some enterprises have different requirements for general data, whether BI tools can provide different solutions for different software and hardware facilities of enterprises. At this point, Fine BI can access large amounts of enterprise data in two ways: FineDirect (direct connection) and FineIndex (built cube ). FineIndex can perform rapid data analysis by extracting and preprocessing data and incrementally updating data. FineDirect provides a SQL -based database direct connection engine, which supports 1 billion to 10 billion data access and real-time data analysis of the big data platform .
2. Data analysis is not only limited to display, but more to exploration
Exploration can be understood in two aspects . First, the front-end data display is "exploration-oriented". How to understand it? The current data display is based on the induction and reorganization of historical data, and lacks guidance for decision-making. The leader got the report and learned that the sales of a certain region rose and fell, and the product market was low in a certain period of time. It can be said that after reading these situations, the leader still does not know what decision to make. He wants to see the comparison with previous years, The report has to pass again. This requires strong interactivity between charts, and users can view them in depth and from multiple angles. For example , Fine BI 's multi-dimensional analysis operations such as data drilling, data slicing and data rotation, and SPA spiral aggregation analysis can perform simple processing on front-end data. The second is "in-depth analysis". The current business intelligence BI lacks the function of data mining. The development of BI tools can be more inclined to data mining and predictive analysis , such as integration with R language, including classification prediction, cluster analysis, association rules, time series patterns and so on.
3. Keep the "lightweight" attribute
At present , the use of BI is gradually biased towards business analysts, and the tools need to be lightweight to reduce the entanglement of technical problems. Domestic BI tools should have more advantages for localized enterprise needs, and have a more accurate understanding of user analysis habits and business logic thinking, so this advantage should be maintained and further developed.
Therefore, as far as the current domestic market is concerned, the development of BI is still in a slow and hot stage, and the future should have unlimited potential.