Tuesday, April 19, 2016
01:15 PM - 02:00 PM
This presentation is about a methodology for estimating the data modeling effort in Data Warehouse projects.
Data modelers are often tasked with reliably estimating the time required to complete data modeling tasks both in agile and traditional waterfall projects. Based on Toyota's formal continuous process improvement framework, Lean, this presentation will show how to leverage two specific Lean tools: Supplier Input Processes Output Customer (SIPOC) and Value Stream Map (VSM) to estimate the duration of data modeling tasks, specifically in a data warehousing environment.
A reliable data modeling work estimate depends on clearly identifying the data modeling output the customer cares about. The SIPOC tool, starting with a holistic view of the customer and the supplier, helps to identify the activities of interest and the required resources in the data modeling workflow. It helps to identify the specific model perspectives and output that are of interest to the customer.
A VSM for the data modeling process identifies the key steps that are required to deliver the expected customer artifacts. Within each step, the activities are identified and the execution sequence is defined. The number of tables and columns are used as input parameters to estimate the effort for each activity. This estimate of the data modeling effort aids the project manager to identify the resources, unearth dependencies and risks, thereby increasing the confidence in the project plan.
In this presentation, two specific customer personas will be discussed, a tech savvy customer, seeking a self-serve environment to create their own reports and a business workflow centric project sponsor, who wants his department to get access to specific reports using a BI tool. These two examples will be used to highlight how the estimation methodology can be customized to meet the needs of both the customers.
Bhaskar Uthanumalliah, a Business Intelligence Architect at Oregon Health Sciences University (OHSU) has more than 17 years' experience in the delivery of data models for Data warehouses. He is engaged in the data architecture and data modeling projects spanning clinical, education and research business units at OHSU. He has extensive knowledge of relational databases, hands on experience with vendor specific implementations (Oracle, SQL Server, Netezza, DB2, Informix etc.) and expertise with data modeling, ETL and BI reporting tools. A seasoned data modeler, he has the full breadth of technical, interpersonal, communication and project management skills that are needed to make BI projects successful.