Data Context: String Theory and Consciousness Bending
I’ve been developing with Microsoft Power BI now for a couple of years. It is a fantastic next-gen tool as compared to SSRS, though not always an outright replacement, at least not yet. With the pace that Microsoft is releasing product updates to Power BI (and other business apps), there is always more to learn and more tools to use. This is both exciting and stressful. One of the recurring difficulties I’ve experienced throughout my time developing data models and reports is fully grasping the extra dimension of data context – both in the model and in the reports. I’m writing this blog to reflect on and help expand the readers understanding of how context works in data modeling and Power BI and to make peace with what I have learned and move on. Data modeling can challenge your abstract thought muscle and bend your consciousness by seemingly pushing the creator into new dimensions of thought.
What do I mean by “data context?”
A good analogy for data context is this – think of a map of the world. The entire map of the world is certainly one grand context, the largest point of reference, the “global” context (pun intended) for this analogy. From here, you can break down the globe into many smaller parts – hemispheres, continents, countries, states, provinces, zip codes, cities, blocks, etc. Now let’s bring a statistic, population, into our global map scenario. From the context of the globe, the largest context, let’s say we have roughly 7.7 people living on Earth right now. As we change context, shifting our focus into some of the smaller parts, like hemispheres, we will see that the number of people living in these smaller parts is smaller than in the global context. For example, let’s look only in the Northern hemisphere and check in with our population statistic. There are 4.9 billion (made up for illustration only) people living in the Northern Hemisphere. Note that this is less than the 7.7 billion people in the global context. If we shift our focus again at yet a smaller context, the country of the U.S.A., there are about 323 million people living there. Notice that 323 million is much smaller than 4.9 or 7.7 billion.
The point I’m making with this thought exercise is that you may have a concrete idea, like population (the count of living people in a specified location), but depending on the context that you examine, you will potentially get vastly different numbers. The context with which you are viewing your data in your reports has exactly the same bearing in the accuracy and usefulness of your reports. With our global map example, this is a familiar visual that we can all relate to and see in our minds. When you start building out data models of far more abstract concepts, it becomes exponentially more difficult to keep your data context straight. I mention String Theory in the title because of the field’s stance that there are additional dimensions that make up our universe, and at times, when building out complex data models, it may feel like you’re expanding your minds ability to use more than the normal 5 dimensions we experience in our daily lives.
When I first began modelling data, I struggled with articulating what I was having difficulty with. Eventually, as I began to understand data context and some of the handy DAX functions like ALL(), ALLEXCEPT(), and ALLSELECTED(), as well as the difference between SUM() and SUMX(), I began to make a shift in my brain – almost as if to add an additional dimension to my conscious thinking. The mental maps of my data models began to take new life and I could more clearly see the relationships that I was building and how the slicers, filters and DAX expressions all work together in Power BI to create what can be very complex algorithms; logic structures that implement business logic that could not be replicated in SSRS. Complex SQL stored procedures used to crunch business logic while slowing down report rendering times in SSRS could be done more simply and elegantly with data modeling. Data visualization and modeling is fascinating and evolving at break-neck speed. The power of these tools is immense, and I feel that businesses are only beginning to tap into their potential! As a culture, I think we need to evolve our understanding of data – in particular, relationships and context. We are in very exciting times to work in technology!