Press "Enter" to skip to content

# Day: November 10, 2023

Imagine two variables, like height and weight, that exhibit a joint distribution. The bivariate normal distribution captures the relationship between these variables, describing how their values tend to cluster around certain means and how they vary together. It’s like a two-dimensional bell curve, where the peak represents the most likely combination of values for both variables.

Click through to learn a bit more about bivariate normal distributions, including ways to plot one and show its density function.

Comments closed

This could be a hot take, but I’m not a fan of pre-reads and will respectfully decline most requests to share content before a meeting.

Before I elaborate on why, let’s start by exploring what pre-reads are, why they often fail, and some more effective alternatives.

I think the viable but difficult alternative is to do what Jeff Bezos did at Amazon: for each meeting, there is a 2-page primer covering all of the relevant context for the meeting. After people are in the room, you distribute the 2-pager and everybody spends 5 minutes reading it first. That serves the intent of the pre-read but there are strict social cues to do the reading, something that does not exist with pre-reads. It also prevents people from going around in circles because they have different subsets of information and don’t realize it.

Of course, this is a challenge to pull off in practice and requires more effort from the standpoint of meeting hosts, but I’m also of the belief that there should be some level of pain involved in scheduling a meeting, as that will cut out many of the “This could have been an e-mail” types of meetings.

Comments closed

As Microsoft Fabric is a brand-new technology (and still in preview – as of 2023-11-09), there are hardly any best practices published and available.
But today, I found some great catches in the Fabric documentation – two decision guides

Click through for more details on those two guides.

Comments closed

Many companies don’t follow the regular Calendar as we know (January 1st – December 31st). They follow their own Financial Calendar (often called Fiscal Calendar) which can start at any time of the year.

Because of this, writing Year-To-Date calculations in DAX for your Tabular Model might seem challenging.

In the step-by-step example, we are working for a company that starts its Financial Year on July 1st.

Read on to see one way to do it. It doesn’t quite solve the problem Olivier brought up, but I’d also make note that having a calendar table with fiscal + calendar year information in it helps remarkably well. It can even handle multiple fiscal year concepts; as an example, a state agency I worked for had a fiscal year on July 1 but the US federal government’s fiscal year begins October 1, so it was just a matter of having StateFiscalYear and FederalFiscalYear columns.

Also, check out Olivier’s new theming, under the Data Cuisine motif.

TimescaleDB is an open-source time-series database extension for PostgreSQL. It is designed to efficiently manage and query time-series data, offering features such as automatic data partitioning, data retention policies, and specialized time-series functions.

This extension provides scalability, improved performance, and seamless integration with PostgreSQL, making it a powerful choice for applications dealing with large volumes of time-stamped data, including IoT, monitoring, and analytics.

Read on to learn how to install it (on Linux), some of the tuning parameters available, and how to create time series hypertables and chunk tables.

Comments closed

Recently a customer was looking to automate some of their SQL Compare checks, but they wanted to do this in a dynamic way, since they needed to do this at scale. Their idea was to not have a SQL Compare project, but build a mapping at the command line.

This post addresses a part of their issue. This will look at how to set up a basic SQL Compare command line.

Read on to see how. This is something I did a while back as well, though it was quite a while ago (say, 8-9 years ago) and required us to do some DLL weirdness back then. Once we got it working, however, it did a good job.

Comments closed