Refer to part 3 of this series for the broad design. Here are the sources in Oracle eBusiness suite for the different dimension and fact tables:

Dimensions:

Facts and Document Dimensions:

Source Data Structures:

The following diagram outlines the source table relationships for the facts above:

The Budget tables’ relationships are given below:

More detailed diagrams can be downloaded from Oracle Metalink.

The descriptions for the chart of account segments are stored in Flexfields within Oracle. The following query, for example, can be used to get the descriptions:

select gcc.segment1 company
, gcc.segment2 account
, gcc.segment3 cost_center
...
, fvc.description company_desc
, fva.description account_desc
, fvcc.description cost_center_desc
...
from gl_code_combinations gcc
,fnd_flex_values_vl fvc
,fnd_flex_values_vl fva
,fnd_flex_values_vl fvcc
where fvc.flex_value_set_id = (select flex_value_set_id from fnd_flex_value_sets where flex_value_set_name like '<YOUR_COMPANY_GL_AFF_COMPANY')
and fvc.flex_value_meaning = gcc.segment1
and fva.flex_value_set_id = (select flex_value_set_id from fnd_flex_value_sets where flex_value_set_name like ' YOUR_COMPANY_GL_AFF_ACCOUNT')
and fva.flex_value_meaning = gcc.segment2
and fvcc.flex_value_set_id = (select flex_value_set_id from fnd_flex_value_sets where flex_value_set_name like ' YOUR_COMPANY _GL_AFF_COST_CENTER')
and fvcc.flex_value_meaning = gcc.segment3
order by 1,2,3,...

ETL Tip:

Loading Fact GL Balances may be tricky if you are not aware that the source table stores balances for all types of amounts - budgets, actual amounts, encumbrances and statistical amounts. Also, it stores the translated amounts if the translation activity has been performed at the close of the fiscal month. As a result of this activity, you may have adjustments posted to certain accounts where you will have actual adjustment amounts posted in the translated currency even though the corresponding local currency amounts are zero.  If you are translating amounts in the Fact GL Balances table on the fly using the reporting tool, be aware of this issue. Although translations can easily be done dynamically, I have found it unwise to do so because of these issues and the complex business rules that often are applied to translations. It is safer to store actual translated values in the fact table itself.

This is the end of this series. I will write about building Sales Data Marts in the next series. Hope you enjoyed these - I am happy to answer any questions you may have or elaborate on anything I have stated in these posts.

Also read Part 1, Part 2 and Part 3 of this series.

Remember the AAMIE of security - Authentication, Authorization, Monitoring, Integration, Encryption.

Download the presentation OBIEE-Secured!

And the cost cutting season has begun. We know the drill: across the board budget cuts, new projects delayed, the organization chart de-layered, voluntary retirement packages and, of course, reduction in force and severance packages.

This is not just an exercise in reigning in escalating costs. The more agile organizations also take such opportunities to align their corporate workforce to the new realities of their business and acquire new employees with the skills that are more suited to their  core competencies and evolving business focus.

This brings into focus the need to retrieve and analyze information about employee performance, learning, compensation, benefits, payroll, time management, and recruiting. As with most modules, there are plenty of pre-built HR Analytic applications like Oracle Daily Hr Intelligence, Oracle BI Applications (Siebel Analytics) – HR Analytic module, Cognos Performance Applications, Business Objects HR Analytic Application, to name the leading few. However, more often than not, you will find that these applications work very well if you haven’t customized your HR application (be it Oracle HRMS or SAP R/3 HR module). They will lack some of the most critical information of your business that you have carefully stored in the customizations of the ERP application as I am experiencing while analyzing an Oracle HRMS implementation at one of my clients. After careful evaluation and cost-benefit analyses, we decided to build a custom data mart, rather than buying a pre-built BI application from Oracle or BO. I will focus this series on building the HR data mart from Oracle eBusiness Suite, based on this recent experience.

Building a HR data mart is considerably more difficult than building other common data marts like a Sales data mart or Finance data mart.  The challenges are many, and not all of them are technical. For example, although HR is the second largest (after Finance) provider of information for internal managers, it has a much tighter budget and a smaller cast of analysts to provide you with good analytic requirements. HR analysts are also very wary of sharing “sensitive” information like salary, benefits, and bonuses about the company’s employees. Thirdly, the skills inventory in HR is rarely taken very seriously and hence HR often struggles to gain the focus and attention of senior management for their BI initiatives. From a data warehousing perspective, extracting HR data is more challenging if auditing and date tracking features are enabled in Oracle (which is true over 80% of the time). Moreover, security in HR systems is far more stringent, given the sensitive data that these systems hold. Security in HR goes beyond standard user or responsibility-level security for system access or even restrictions on data visibility based on the organization chart. I will write more about it in a later post, when we talk about the overall design of the data mart.

The HR data mart, particularly in the context of Oracle eBusiness suite has a rather large functional footprint. You need to understand several functional areas / modules to be able to design a data mart that is effective in meeting the reporting / analytic goals of your users. Of particular interest are: HRMS (Core – Employees, Recruiting, etc), Advanced Benefits, Incentive Compensation, iRecuitment, Payroll, Time and Labor, Performance Management, Learning Management and Workforce Scheduling. Although it is not common for a single company to require analytics for all these subject areas, a subset of this (HRMS, Performance Management, and Payroll) is very typical. Most companies have specialized HR reporting needs, for example, on immigration – that are not covered readily in a standard implementation of Oracle HRMS. More on that as we get to the analytics.

Before anything else, here is a quick map of how this series on Human Resources Analytics will be laid out:

Part 1: Overview of Oracle HRMS
Part 2: Key analytics for HR
Part 3: HR data mart design and considerations
Part 4: Source system, ETL tips

Stay tuned…