The methodology that we use in the implementation of Business Intelligence projects is based on an agile approach that can minimize the costs and “time to market”.
The cornerstones of this methodology are:
- Agile Approach
- Just Enough Design Upfront, a slim preliminary planning which is able to achieve the project’s objectives.
- Just Enough Documentation, essential and effective documentation.
- Process automation. Where possible, portions of repetitive work or of work related to fixed patterns are delegated to automated procedures. This results in a significant reduction in implementation time.
- “Data Lake” Enabled. The methodology is compatible with new technologies and trends that see the traditional Business Intelligence systems as increasingly more integrated with “big data” systems (Hadoop in particular), with which they form the so-called “data lake”.
- Evolved dimensional model. Starting from a dimensional type of modelling for the realization of the data warehouse, we have added effective and improved design techniques.
The DataSkills methodology starts with a preliminary analysis of both the business environment and the problem, which we split into self-consistent “work units” that are easier to implement. The process is developed via an iterative approach that comprises of the following steps:
- Work unit analysis
- Multi-level modelling:
- Logical modeling
- Physical modeling
- Semantic modeling
- Prototyping, able to generate a semi-finished product that can show the customer the solution to the business problem.
- Verification by the key users. The verification phases can be more than one for each work unit, up to the final verification which occurs once the work unit is complete.
- Development completion. This phase takes place in various steps, interspersed with verification stages.
- Deployment. This final phase involves, in addition to the final release, a training process for the company’s end users and IT staff.
Leave A Comment