What we're about
Upcoming events (3)
Let’s see, what is the least exciting lecture subject we can conjure up? • The mating habits of the lesser spotted fluke worm? • 101 uses for worn out sandpaper that don’t involve glue? • How not to wind an automatic watch? None of the above because a lecture on p values is clearly far, far less exciting. And, to ensure maximum boredom, this talk is not even about the statistical tests that generate p values, (which we all know are terminally boring in themselves) this lecture is simply about the generalised meaning of that p value thing itself. Which is really silly because we know that a p value is designed to prove (or disprove) … well… whatever it was that you were trying to prove with the statistical test in the first place. And this does bring us to the crux of the matter and the real reason for the talk. One of the many things that p values do not do is to prove or disprove anything. Ever. So we have this very bizarre situation where a metric (the p value) is actually stonkingly useful, often quoted and almost universally misunderstood. This lecture will attempt to explain, in non-mathematical terms, what p values do, why they work, to some extent how they work (without going into the maths) and why they are so amazingly useful when you really do understand what they are actually telling you. PS no maths!
This case study demonstrates and explains different types of relationships in Power BI data models - in particular bi-directional, many-to-many and virtual relationships. We'll solve the problem in Power BI by first building a data model with bidirectional and ambiguous relationships before realising our mistakes and improving on it the help of some DAX, the calculation language of Power BI. We solve the "nearest neighbours" problem -often we need to compare the performance of something, for example an organisation or equity, against a set of similar things. The challenge arises since A may have neighbours of B and C and B may have neighbours of C and D but not A - so we can't simply group. This is the first of a series of typical Power BI data modelling challenges "from the trenches" - in subsequent sessions we will be looking at peak values, and later at gross and net values. This session will be presented by Mark Wilcock.
Relationships are a core element of Power BI. In this introductory session we will explore the basics of relationships: how they work and why we use them. In doing so, we will cover: • One-to-One; • One-to-Many; and, • Many-to-Many We will also explore how relationships improve the user experience, resolve internal inconsistencies and reduce repetition of data. Resources (worked examples, datasets etc) are at https://bit.ly/33zWQtp in the 'Basic relationships in Power BI' folder. If you would like a sneak preview, Chris has already recorded a video of the content at https://youtu.be/cuDaajadNmw