Remapping

When I first learned about neural networks several years ago, I found it odd (but perhaps also interesting) that they taught it as a neuroscience analogy. I'm not sure if I came into data science on totally the wrong track (mathematics as an undergraduate for the record), but I didn't find this analogy helpful because I didn't take a biology class after high school.

This does, however, shine a light on an interesting discussion around how we learn new things. If you've ever found yourself learning something in a totally different way than your colleagues, you're not alone! Analogies for learning new things work a bit like targeted marketing in that we have to know our audience before introducing the comparison. Otherwise, it totally misses the mark. When I'm creating analogies for my LinkedIn Learning courses, for example, I try to use examples that either make sense to someone living in the everyday world or are part of a topic that I would expect someone taking the class to know beforehand (introductory statistics for example).

Our brains work in fascinating ways. For all the discussion around paragraphs of text generated by models like ChatGPT, it's worth noting that humans wrote the initial text to start with. Our output is only as good as the input we put into it!

Actual Maps

Along the lines of mapping our brains, we can also map geopolitical boundaries or even our own maps (for example, if your organization has a boundaries map for sales regions). If your organization doesn't have custom maps, the great news is that there are lots of publicly available maps you can use instead. If you live in the US, the best place to find the latest maps of states, counties, and congressional districts the TIGER/Line Shapefiles section of the US Census website. After I worked a lot with maps, I realized I needed to learn more about how to edit them. This is where the online map editing tool mapshaper came in. An example of how to use mapshaper in tandem with Power BI is the focus of a recent Power BI Weekly video, where I show how to create a map of counties using the clipping options for map editing to only show the counties in three US States.

As the mapshaper website description entails, maps themselves are shapes representing geographic boundaries determined by latitude and longitude points that create their vertices.

Mapshaper takes existing shapefiles in formats like GeoJSON and TopoJSON and lets us edit them. We can then export them in these JSON formats to use in our own tools like Power BI.

DAX

Another topic that was a bit mind-bending when I first encountered it was the DAX language. DAX (Data Analysis Expressions) is the language that we can use to write queries to get data tables from multidimensional database sources like SQL Server Analysis Services. We can also use it to create tables, columns, and measures in tools like Power BI.

Rules of DAX

DAX is not technically a programming language. However, it still comes with all the frustrations and ultimately rewards of using a programming language like Python. It has quirks and technicalities that we need to know first in order to navigate our way around using it effectively. In the case of DAX, once I knew the order of operation in functions like CALCULATE, I could make a lot more progress in terms of building the DAX measures that I wanted that worked in the way I expected them to. DAX formulas work using the following logic.

1. Filter the data
2. Perform an aggregation on the filtered data in Step 1

DAX can also be deceiving. While it shares many of the same functions as Excel, it operates on tables and columns instead of cells. This makes the syntax of it look deceiving easy compared to the logic of creating measures that work because the syntax of the formulas often reverses the order of the logic of the steps we see above.

Houston Power BI User Group

In the last year, the DAX functions below debuted in Power BI. If you live in Houston, this coming Thursday I'll be doing a presentation on these nine new DAX functions at the Microsoft offices on the west side of the city.

Here's the link to sign up for the meeting!

February 2023 Greater Houston Power BI User Group Meeting, Thu, Feb 16, 2023, 4:00 PM | Meetup

Coming up!

I'm prepping to record another quarter of the Power BI Weekly videos right now! My time series modeling course in the LinkedIn Learning library is currently in the editing process for the amazing people who work their magic on the raw videos in studios outside of Santa Barbara, California. I'll share in this newsletter when the course comes out, stay tuned!

-HW