Business Intelligence (BI) Testing with Sample Test Cases

The procedure in which gathering, cleaning, integrating, analysing, and sharing data is done to
determine actional experiences that drive business development is known as Business Intelligence
(BI). Business Intelligence Testing checks the organizing information, ETL process, BI reports and
guarantees the execution is right. BI Testing guarantees information credibility and exactness of
experiences got from the BI process. Here, we will discuss the following points:
1. Events in Business Intelligence.
2. Testing Sequence of Business Intelligence.
3. Stages of Business Intelligence.
4. Business Intelligence Test Cases.
5. Strategy of Business Intelligence Testing.

Events in Business Intelligence - The flow of events of business intelligence are:

1. Client transactional data (Relational database, or OLTP) Flat file database, records
of data: It is a type of data that is recorded from the transactions of the clients. These
data are stored in a form of records in a flat-file database. A flat-file database is a kind of
database that has a uniform format and does not consist of any kind of indexing or
structure. It is a database in one table format. Mostly it is saved in CSV format.

2. ETL processes: The method of extracting data from numerous source systems, then
transforming the data and loading it in the data warehouse is known as the ETL process.
The full form of ETL is Extract, Transform and Load. In this step, the data from the flat file
database is extracted. The transformation is important as it involves various business
rules and also has some risk of miscalculations. This is one of the most vital steps.

3. Data Warehouse: The process of handling and collecting data from different sources to
provide a significant understanding of business data is the data warehouse. It builds the
data analysis and reporting. It is a process of transforming the data into information and
making it available for the user in a periodic way.

4. Data Mart: It is a subset of a data warehouse where it primarily focuses on a single

subject line of business. Datamart collects data from a few sources and is very much
more flexible than a data warehouse. The understanding of data is built in this step.

5. OLAP to generate significant BI insights: This is a step of computation that enables

the user to extract the data selectively and query data to understand the different points of
view. The full form of OLAP is Online analytical processing. All the data in OLAP are pre-
summarized data thus it takes very less query time to execute. Finding a correlation
between various data is done here.

For Example, Recommended items on e-commerce sites, Recommended videos on YouTube, etc.
The technologies/ systems that are commonly used for business intelligence are:
1. MIS: Management Information System.
2. OLAP: Online Analytical Processing and Multidimensional Analysis.
3. CRM: Customer Relationship Management.
4. EIS: Executive Information System.
Stages of Business Intelligence - The BI testing has two stages:

Stage 1: Processing and storing the data:

 Source data: The data in the source system could have some problems regarding how
the data has been entered. The BI engineers could not manage the source data, and it
leads to a possibility to influence the source report. For that reason, it is vital to validate
the integrity of the source data to guarantee accuracy.
 ETL: Once the data has been collected from the source system, it is then changed over
and transferred to the data warehouse. This change is essential since it includes business
rules, which is additionally why there is a high opportunity for errors, and miscalculations
at this stage.
 Data Warehouse: Regardless of whether no errors are found in the source testing, the
data warehouse could be the issue. There is the probability that a few orders could be
missed in the data warehouse prompting these issues. The data for these orders have
been coincidentally lost.

Stage 2: Testing of BI:

 Meta-data layers: It provides high-level objects with simple access to business clients.
The data here is collected from databases and a transformation of data is considered
here.
 Reports: Each BI report is comprised of SQL queries, prompts, and filters. Issues could
emerge in any of these things because of specialized or development errors. Creating
these reports is a significant improvement scheme, which is the reason it should be tested
to guarantee that all data is exact.

 Dashboards: The dashboards in BI testing consolidate a few reports with various data

and graphs. These two could conceivably be associated. As a rule, the dashboards are
the last instructive pieces utilized by businesses, that’s why testing it is necessary.

Strategy of Business Intelligence Testing

The various strategies of Business Intelligence testing are:

 Organizing and planning for a test.
 The strategy and technique for the test.
 Design of text (For example most of the test cases are concentrated towards query rather
than simple text). This is the most important difference between ETL/Data Warehouse
projects and the traditional test project.
 Test Implementation
 Reporting of limitations and conclusion of the project.

Testing Sequence of Business Intelligence - The testing sequence of Business Intelligence is:

1. Verify the Source data: Business Data generally doesn’t come from one source and in a single
format. Ensure that the source and the kind of data that it sends match. Basic validation is done here.

2. Verify the transformation of data: This is the place where raw data is processed into business-
explicit data. The source and destination data types should be equal. The primary key, foreign key,
default, and null value must be untouched. ACID properties of source and destination data types must
be verified.

3. Verify the data loading: The data that is being loaded and tested by the scripts are added for the
ETL testing. The data storage system should be verified for the accompanying:
  Performance: For complex systems, connections emerge between different parts of the
systems forming various co-relations.  Though it is good for data analytics still a lot of time
is required to retrieve data. Therefore, performance testing is the major factor.
 Scalability: Data is increasing day by day. Therefore, testing of the data is required to
decide if the current implementation can handle the data of the increasing business
volume or not.

4. BI report testing: This is the thing that is viewed as Business Intelligence. Note that if the previous
layers are broken, the reports won’t ever be exact, reliable, or rapid.
The important points are:
 The utilization of the reports that are created for business.
 The parameters that are mentioned in the reports should be modified and customized like
sorting, grouping, categorizing, etc.
 The presence of the actual report i.e., the documentation.
 The comparing use of the application is to be included in an end-to-end test if the
elements of business Intelligence are integrated together.

Business Intelligence Test Cases - The BI test cases are:

1. ETL Verification:
 Confirm that data is planned accurately from source to target system
 Confirm that all tables and their fields are duplicated from source to target
 Confirm that invalid fields for example null values are not occupied
 Confirm that arrangement of keys to be auto produced and are made appropriately in the
target system
 Confirm that data is neither confused nor shortened
 Confirm that there is no duplicate information in the target system
 Confirm that data type and arrangement in the target system are accurate.
 Confirm that the accuracy of data in numeric fields is precise
 Confirm that changes are applied accurately
 Confirm that exception handling is powerful.

2. Staging data:
 After executing filter rules the record count between the staging tables and target tables
are equal.
 Insert a record that is not present in the target table in case of the supplied key
combination.
 If the records are pre-loaded, they can’t be duplicated or sent to target tables.
 Remove the records in the target tables coherently.
 Values are loaded in the process tables
 Values are loaded in the reference tables.
 At the time when day_02 data loads, update the records for the key.

3. BI Data Loading:
 Verify that there is no issue in accessing the data and the source and destination
databases are appropriately connected.
 Verify the truncate option if it is working properly or not for the loaded data.
 While loading the data, verify the session’s performance.
 Search for mistakes that aren’t deadly.
 If the child’s process fails, make sure that the parent’s process will fail too.
 Make sure that the log is updated regularly.
 Ensure the planning and work process parameters are right.
 Make sure that the quantity of tables in the source and target system is indistinguishable.
 Compare the properties of the stage table with the destination tables. They should be
equivalent.
4. Reports of BI:
 Show date and time.
 Accuracy of decimal is required for significant figures.
 On each page, the number of rows and columns should be displayed.
 In the report, there are no free characteristics.
 Null values shown for both attributes and important figures should be clear.
 Check if the text search function is case sensitive or not.