The general process of data analysis is data collection → data transmission → data modeling → data statistics/analysis/mining → data visualization/feedback. Therefore, data collection is the basic and source.

The DATA acquisition SDK generally contains two basic functions:

• Collect data through buried points • Transfer the collected data to the specified server

The data acquisition SDK is required to ensure the accuracy, integrity and timeliness of data to the maximum no matter it is collecting data or transmitting data, which requires the data acquisition SDK to deal with many details, such as: User identity, network policy, data cache policy, data synchronization policy, data accuracy and data security, etc.

At present, there are three main burying points in the industry:

• Code buried • Full buried • Visual buried code buried refers to an application that integrates the buried SDK, initializes the buried SDK at startup, and then invokes the methods provided by the buried SDK to trigger an event when an event occurs.

Code burying is both “original” and “universal” because it has a number of advantages:

• Can accurately control the location of buried points

• More convenient and flexible customization of events and properties

• Richer business-related data can be collected

• Can meet more refined analysis needs

Of course, there are drawbacks to code burying:

• The cost of up-front burying is relatively high • If the analysis requirements or event design changes, the application will need to modify the burying concurrent version

Full burial point, also called no burial point, no code burial point, no trace burial point, automatic burial point, etc. Full buried point means that all or most of the behavioral data of users can be automatically collected in advance without the application development engineer writing code or only writing a small amount of code, and then the required behavioral data can be screened out and analyzed according to the actual business analysis requirements.

The events that can be collected at the full buried point are: • Application startup event ($AppStart) • Application exit event ($AppEnd) • Page browse event ($AppViewScreen) • Control click event ($AppClick) • Application crash event ($AppCrashed)

Full burying has several advantages: • Relatively low upfront burying cost • No application needs to modify the burying concurrent version if analysis requirements or event design changes • Effective solution to the “historical data backtracking” problem

At the same time, full burial sites also have some disadvantages:

  • For technical reasons, it is difficult to achieve full coverage for complex operations such as scaling, scrolling, and so on
  • Service related data cannot be automatically collected
  • Unable to meet the needs of more refined analysis
  • Various compatibility issues (such as compatibility between Android and iOS, compatibility between different system versions, compatibility between different ROM vendors of the same system version, etc.)

Visual burying point, also known as circle selection, refers to burying point through visual way. There are generally two application scenarios for visualization buried points:

  • By default, no burials are made, and then visually specify which controls to bury (specify burials)
  • By default, all are buried, and then visually specify which controls are not buried (excluded from buried).

To learn more, you can download the white Paper on iOS Full burying Point technology for free. This white paper mainly introduces full burying point as the core, and some content is also applicable to code burying point.