• Seilevel Team

    Here’s the Team

    Welcome to our Seilevel Photo Op.

How Do You Approach an Analytics Project?

Business Analytics projects – where do you begin? In a previous post, Joy mentioned using decisions to prioritize requirements for business analytics projects. Beginning your project by identifying these decisions, uncovering business problems and mapping both to business objectives and the product concept will allow you to build out documentation. Many analytics projects are greenfield projects, and do not have any structure around them. At Seilevel, we understand the importance of working toward requirements by utilizing RML and modeling important information to check for completeness and correctness. Below is a high-level template to help you get started on your analytics project, after you have identified your business decisions.

So, your decisions have been elicited. What next?

1. The important next step is to draft and verify a Business Objectives Model. The Product Concept within this model should map to the decisions that need to be made from the analytics. Implementing the “features” of the analytics project will allow the stakeholder to make these identified decisions.

2.After the features have been identified, organize them onto a Feature Tree to provide structure and a clear picture of relationship between all. The Feature Tree also makes elicitation easy – bringing in your one page of organized features to discuss will help stakeholders visualize the features and help uncover any missing branches in the tree.

3. Having identified and organized the features, produce an Objective Chain to help further prioritize your features based on value. The Objective Chain will allow you to rank the importance of the decisions/features – which may be especially useful when purchasing or building an analytics engine.

4. Documenting Use Cases for each decision will help give context to the necessity and reasoning behind including it. Use Cases also transition nicely into requirements.

5. Systems and Data: In addition to modeling the usage of and reasoning for business intelligence and analytics, identifying the data pieces and systems involved will help you document non-functional requirements. In my experience, creating the following models to elicit low level detail simplifies the requirements writing process.

Ecosystem Map – capture all interlocking systems from which data pieces and algorithms may be extracted.

Business Data Diagram – identify all data pieces that will be utilized to perform analytics, and understand the relationship between each.

Data Flow Diagram – document how the data will be transformed through analysis, which actors will be receiving the data, and from which data storage location the data will be pulled for transformation.

All of the models listed above can be found via the links above or in Visual Models for Software Requirements, and on our website www.Seilevel.com.

In my next post, I will be focusing on integrating a COTS Analytics Tool – an integral part of any new analytics project.

Business Intelligence and Analytics projects are trending across all business sectors – how has your team approached these projects?

Leave a Reply

Your email address will not be published. Required fields are marked *