Job Role: Business Intelligence Analyst

A business intelligence analyst has a key role in a company. This job is related to integral software and data assets that help firms to chart a course forward and make decisions about operations.

In the broadest sense, the business intelligence analyst is responsible for working with business intelligence data, creating insights and establishing ways to utilize these data assets within the company. (What exactly is business intelligence? Learn more about it in An Introduction to Business Intelligence.)

This overall job role has various components. Business intelligence analysts may work gathering and documenting technical requirements, and showing how reporting technologies and analytics tools factor into business operations.

They will typically use SQL database tools and other technologies to create reporting solutions and actively mold different sorts of data to support insight generation.

Another key component is to support the end user — for example, a business intelligence analyst may assist in creating or enhancing a particular kind of visual dashboard, or might even train users on it after it’s created.

For more, let’s take a look at an average ad for this professional role.

More on the “User Side?”

One way to try to delineate what a business intelligence analyst does is to contrast the role to others that might be more fully technical in nature.

Here’s what Ganes Kesari, co-founder and Head of Analytics at Gramener, has to say:

BI Analysts work on identifying insights from data and converting them into visual stories. They use self-service data discovery tools such as Tableau, Power BI. Many of them have good SQL skills and moderate programming background. However, don’t count on them for heavy backend development, which is the domain of ML Engineers.

BI Analysts are often aligned with the IT team, but work closely with business users. They understand the functional needs, are able to plumb the data, put together queries, and create ad-hoc BI dashboards.

Getting Technical with Advanced BI Analysis

Although platforms like Tableau are popular in the business intelligence world, some analysts go beyond these built platforms to utilize raw coding in Python and assorted languages, or build numerical programs “from scratch.”

“I use Python-based tools such as NumPy and Matplotlib when the more convenient yet basic tools don’t have the function I need,” says data scientist Bruce Kuo at Codementor, while conceding that Tableau and Looker are usually his first stops. “Also, BI engineers focus more on data processing logic. The goal is to provide better quality data for business analysts.”

Kuo talked about using NumPy, an array library for Python, in BI work.

“When processing,” he says, “I may use NumPy operations, but not Matplotlib as I don’t need to create visualizations. In most cases, I can generate insights from simple rules and assumptions if I consider the problem carefully. It is generally unnecessary to engage with machine learning programs, except for clustering cases like tagging groups of users by user behavior, because user behavior is difficult to define by rules.”

Kuo further explains that NumPy is helpful in importing inputs in a particular format.

“I use NumPy for BI when the visualization library or package needs NumPy type inputs,” he says. “For example, we need a ‘USB shape’ to work with the ‘USB port.’ In this case, the ‘USB shape’ is NumPy and the ‘USB port’ is the visualization package. Data must be processed to the right format before being imported to visualization tools … For me, NumPy doesn’t act as storage, rather it is more of a computing concept.

For example, you can do matrix operations with NumPy, which is fundamental to machine learning. In engineering, NumPy implements various optimization tricks to help improve performance.”

Then there’s also the power of machine learning, which Kuo says can help out in creating more sophisticated models than we can “think about” on paper without automated analysis.

“Before machine learning, people tried to define and separate users by rules,” Kuo says. “For example, we know 30-40 years old users tend to buy beer. However, sometimes simple 1-5 factors are not enough to describe users correctly. The number of rule combinations is simply too large for the human brain to compute. So how do we solve it? This is how machine learning began. ML can consider many factors efficiently. As a data scientist, my job then becomes defining problems, metrics, or experiments carefully.”

The Future of BI Analysis

Is the business intelligence world accepting machine learning as a future innovative disruption? Yes, according to Dave Mariani, chief strategy officer at AtScale.

“Given the ease of use improvements of machine learning platforms and the introduction of data wrangling tools, there’s a new category of BI analysts emerging, sometimes called the citizen data scientist,” Mariani says. “These advanced BI analysts are becoming more data savvy and are beginning to use some of the machine learning platforms to generate predictions and create features.” (For more, check out The Role of Citizen Data Scientists in the Big Data World.)

Mariani’s assessment fits with what is often called the “self-service model” — the idea, with cloud services up to and including machine learning platforms — is that end users who are becoming more technically proficient start to use these platforms on their own rather than relying on “IT people” to do queries, build tables or generate insights.

This idea also sort of combines all of what the BI analyst role is about. It’s getting the data, working with the data, and delivering value that will constitute “business intelligence” in all its splendor and diversity. How does a company innovate its product lineup? What specific categories of customers represent funnel activity? Why change a brand or logo?

BI analysts help to answer all of these compelling questions, and more, through data science work.