A Typical BI Process
A typical BI process includes 4 steps:
- Business Analysis: How to make money in a business? Core processes and key decision areas; goal and objectives; strategies, tactics and operations.
- Data Analysis: Preparing data for analysis is more than half the battle; data sourcing; data pre-processing, exploratory data analysis (EDA) and summary statistics, data issues and treatments, extract-transform-load (ETL).
- Data Modeling: A data model reflects the business model in a digital world; data models are a shortcut to understanding the business model; entity relationship diagram (ERD), using primary key and foreign key to create one-to-many relationships between dimension tables and fact table.
- Data Visualization: Storytelling with data, producing interactive reports and dynamic dashboards with a 360-degree view of a business; dimensions and measures (KPIs), rollup and drilldown, slice and dice, sort and rank, conditional formatting and traffic lights; visualize trends and drivers, present conclusions and recommendations with real impact on business (e.g. revenue growth, cost reduction, efficiency gain, quality improvement).
BI Roles and Responsibilities
Career opportunities for BI roles have grown exponentially to meet the increasing demands of digitally transformed industries, businesses, functions, and processes. Organizations are looking to fill 4 major BI roles in a BI team with overlapping responsibilities: BI Analyst (junior) and BI Consultant (senior) who spend more time in step 1 and 4 of the BI process, BI Developer (junior) and BI Engineer (senior) who spend more time in step 2 and 3 of the BI process. When there isn’t a BI team and separation of duties, you will need to do all 4 steps and cover the entire BI process.
BI teams combine business analysis, data analysis, data modeling, and data visualization into one end-to-end BI process to deliver various BI services. For example:
- Building pipelines. Define KPIs. Develop Dashboards. Automate Processes.
- Determining when to use what charts to visualize measures by dimensions: line (trend of time series), pie (part-to-whole relationship),column or bar (comparison), histogram (distribution), scatterplot (correlation).
- Rollup and drilldown, slice and dice, sort and rank, present findings, recommend actions to monetize an opportunity (before window closes) or fix a problem (before it gets bigger).
- Presenting a 360-degree view of a business, from different angles (product, marketing, sales, operations, finance) and at different levels (analysts, managers, directors, VPs, C-Suite).