Intelligent testing is the process of not just collecting raw/business data, but also transforming it into data that has relevance, use and meaningfulness. The raw data is the record of daily organisational transactions such as interactions with customers, finance administration, employee management and so on. Data is used for reporting, analysis, predictive analysis, data mining, data quality as well as interpretation.
Creating a data warehouse enables the company to engage in data consolidation from several sources and separate analysis from transaction workload. Data is turned into top quality information to meet enterprise reporting needs for different users. Additionally intelligence testing includes examining source data, data mart, reports, target data, ETL, integration, reporting services, analysis service, cubes, database engines, measure groups, reports testing, facts dimensions, drill down drill through, build deployment and so many more.
Stages of Intelligent testing
Numerous data sources can be listed or non listed ones. These serve as an input to the testing software to enable the taking of strategic decisions by the end user.
Staging/ Data Collection
This involves creating source data inputs to store in the data warehouse where data is loaded using logic and integration services from source data. It is data in its original form, mostly with changes needed for report generation.
Data warehouse/ Data Integration
This is when a database engine gets created. It is the final object that serves as a basis for cube creation for analysis purposes. Object types that data are organised into include facts and dimensions.
Analysis/ Data Storage
This is the analysis phase for intelligence testing. Most commonly, 3D data modelling techniques are used when data is analysed in more than just two dimensions. The facts are analysed in association with parameters which are known as measure/measure groups. These are combinations of more than a single related measure. Greater insights into business decision making are possible with variations in parameters.
Reports/ Data Presentation
The data is then represented on user friendly, intuitively designed interface to create reports as per business needs. It is the end product and the final outcome that gets developed. It is drilled down and drilled through, in association with the scenario required.
Value of Testing
Effective testing integration in the implementation process creates confidence and trust among business users, in association with the BI data generated. The typical challenges faced in intelligent testing of raw data are as follows. There’s data volume, complexity and variety first of all. This creates a problem, as disparate data sources cause data anomalies. Data loss through data integration processes and handshaking between sources is there. Additionally, intelligent testing is time consuming and no audit trails, or methodology result in high cost of quality. Moreover, specialised skills are required to execute data verification and validation.
To ensure data is complete, accurate, secure, consistent and reliable through the lifecycle, it is critical to check all aspects of the data entry point into the intelligence architecture and not just at the end through dashboards or reports.
Importance of Testing Strategy from Pathwwway Gaming
Goal of intelligent testing is the attainment of credible data. Data credibility is achieved by making the testing cycle efficient. A detailed, comprehensive and holistic testing strategy creates more advanced test cycles, covering test planning for every stage, each time data movement takes place. Stakeholders like developers, business users, QA team, infrastructure team, business analysts and more from key areas need to focus on:
- Testing Scope
- Availability of test data
- Data quality and performance judgement criteria
- Test environment setup
Data collection, integration, storage and presentation are roughly the stages of intelligent testing. Data should cover all critical business scenarios.
Primary aim of data completeness is all of the data extracted should be loaded at target. At the time of data acquisition, the factors that need to be considered are time boundaries of selected data, various data sources and other special cases. Validation of the data required and availability of data sources from which data needs to be extracted are important Data profiling activity helps in the understanding of data. Identification of data problems will lower cost of fixing later in the development cycle.
Testing within the data integration phase is critical as data transformation occurs at this stage. Business requirements undergo transformation logic. Validating data model, involves checking the data structure within specifications. Data dictionaries need to be reviewed. Following this the source needs to be validated to target mapping and data storage is when the analysis takes place.
Validating data loads are based on time intervals. Initial and subsequent load testing comes with performance and scalability aspect ensuring the system is within performance limits and can sustain data growth. Verification of appropriate parallel execution and precedence during the ETL process is critical as it impacts performance and scalability of the system. Ensuring data history is based on business requirements is important. Error logging, recovery from failure points and exception handling need to be verified.
Data presentation is the final step of the testing cycle. Additionally, there is a GUI to test data.
End to end testing forms the key here. Along with data reconciliation discrepancies, issues like deadlocks or resource contentions need to be studied. End to end runs ensure data quality and performance acceptance criteria are reached.
Intelligent testing of raw data serves to ensure end to end traceability to reports through source extraction. Time savings through automated regression testing, defect or bug tracking involving stakeholders and improved testing time through reusability matter.
Important aspects that Pathwwway Gaming believe need to be considered are business logic and complex architecture, large volumes and data complexity. Data selecting from varied data sources is beneficial. However, selecting information across wide and varied sources can be challenging. Additionally the lack of consistency and constant expansion of data in a data warehouse matter. Changes in business requirements and alterations in source databases, all come into play.
Intelligent data testing is all about accuracy, precision, consistency and reliability. Testing in stages, the main aim is to ensure that raw data yields meaningful insights that can be used to ensure the advancement of the organisation is assured.