Power BI for Business Analysts – Where do I start?

As a business analyst, what exactly should we know to be able to convert the requirements of end users into Power BI reports? What should we learn to be good at using Power BI?

These were questions directed at me by two business analysts at a recent meetup. They were getting into Power BI in their organization, and were given the task of converting requirements of several business stakeholders into analytical reports for BI purposes. I promised them a few links to help them learn. However, links without context would not be fair. Hence this post, with context and guidance on how business analysts can get started with analytics.

When it comes to business intelligence, the traditional role of business analysts in software houses cannot be used to define the business analyst who is deemed to work with data. They have a whole new, dynamic and a more interesting role to play. In a nutshell, this business analyst would need to understand the problem that the stakeholder or end users have and come up with a solution to solve that problem with data. For example, the end user, the director of sales sees that targets for sales are not being achieved nation-wide. They need to be able to see where things are going wrong, and what may be causing the problem. To solve the problem, the business analyst (BA) will have to collect the relevant data, and use a tool to churn out insights from this data easily and efficiently, and then present it to the end user for business direction.

Typically a BA would need to:

  1. Understand the data. Know what and where the right data lives, then, most importantly, understand the data: What information is stored, why it is stored the way it is stored, what are the rules associated with the data and so forth. Basically, the relevant data must be understood as if it were their own.
  2. Prep the data, and mash it up into a data model that represents the business process(es). Data needs to be pulled in from their multiple sources, cleaned, structured and stored in such a way that is intuitive for business users to understand. This is essentially the data-discovery aspect of things.
  3. Derive insights from the data by putting up visual reports and dashboards to tell stories of what is going on with the business. This is essentially the self-service aspect of things.

One of the first things you would want to do as a BA, is get to know the concepts of Power BI, and how it can be used as a business analyst’s tool. The best place to start would be Power BI Guided Learning: Getting Started.

Then comes understanding the data, which is something a business intelligence tool cannot help much with. This would be a task that mostly involves discussions and research involving the owners of the various systems, and finally documenting the findings of the source data. Power BI can come in handy at a later stage of this exercise when you examine and evaluate the data.

Prep Data

Once all the questions are asked and responses documented, you get into the next step of prepping and modeling the data. The easiest would be to bring in all the required data into a flat structure; one great table that has all you want to build visualizations and reports. However, this approach would not work beyond a certain level. When you bring in more than one business process such as sales invoicing and sales planning (targets) together, due to non-separation of common data elements, comparing sales against targets would be difficult, especially when granularity comes into play.

Learning a data modeling technique, such as dimensional modeling will greatly help with structuring the data into an easily consumable piece. Dimensional modeling is a relatively simple technique to learn. Of course it becomes a little complex as you get into the nitty-gritties of some of the concepts, but in most cases, at least in the context of business analysts it wouldn’t come to that.

Model data

A star schema would not be quite enough if you do not build upon it and create the measures you want for your analysis, for example you may already have a sales value measure which comes from the star schema, but if you would want to make more sense of it you would probably require a Last Year Same Period Sales measure and a Sales Growth % measure to better illustrate the state of current sales. Creating these measure on the model that you created would require a little more technical know-how. However, assuming that a business analyst already works with Excel including its functions and formulas, learning the DAX (Data Analysis eXpressions) language to formulate measures won’t be too hard. When it comes to self-service, your data model does not have to be the best representation of the business process(es), but the better it is modeled, the easier it is to derive insights.

  • Start with Power BI Guided Learning: Modeling to learn how to build relationships and enhance the star schema with analytics
  • Then, spruce up your technical prowess with Power BI Guided Learning: Introduction to DAX

Visualize data

Once the data model is done, the next step is to start on self-service analysis of the data. Self-service involves asking questions off the data, and depicting it on a canvas, and creating a story out of it. Not everyone can ask questions of the data. You need to know the business processes and the data in order to do that. And as a business analyst it becomes your job to know the data. This would allow you to ask the right questions off the data, in the right context and build a story out of the answers that you get.

  • Start with Power BI Guided Learning: Visualizations to get an understanding of Power BI visualizations.
  • Use this as a guide when determining which type of visual to use for what purpose: Extreme Presentations: Chart Suggestions – A Thought-Starter
  • Then go further on exploring data with Power BI Guided Learning: Exploring Data
  • Finally, these real-life stories can be used to inspire and provide guidance as to how you can tell stories: Power BI: Data Stories

One thing about visualizing data is that it is part-science and part-art, hence additional reading, practice and experimenting is encouraged.

Publish and share

Once you’ve built your story, which would be in the form of dashboards and reports in Power BI, you need ensure they are shared with the relevant stakeholders, feedback taken, and the solution refined until proper business value is seen.

Further learning

Once you’ve gone through the links above it would be a good idea to take some time and complete the following EdX course, which would give you a holistic and practical learning of Power BI: Analyzing and Visualizing Data with Power BI and then a read of Reza Rad’s (b|t) online book on Power BI:  Power BI From Rookie to Rockstar which will give you a real world feel of how you could you use Power BI. You might need to invest a few weekends for this, but it’s an endeavor that is quite worth it.

Finally, to top it all off, consider certifying yourself. Being certified gives you the validation that you are technically thorough in the subject area. The MCSA: BI Reporting cert is the ideal certification for a business analyst to to showcase their expertise in business intelligence and analytics; it just takes two exams to gain the certification.

Conclusion

Business analysts are expected to be technically savvy, especially when it comes to working with data. Just as you would expect a secretary to know office productivity tools nowadays, as compared to 10 years ago when it was only desirable for them to be Microsoft Office-savvy, it is expected that a business analyst too be tech-savvy about data. Knowing how to work with data, what tools to use, how to use the tools, the concepts behind working with data, and finally an endorsement of this knowledge will be key to excel in this field.

Note to all Power BI experts: Better and additional links to free learning resources would be helpful to keep this post up-to-date

Advertisements