What exactly is a cascading filter?
Ok, let’s say you are doing some analysis on a dashboard and you need to some deep digging. You have a hypothesis that something interesting might be going on with a particular Product Category so you filter on that category.
Now, in order to limit the data to be shown only for Products within that Product Category, you should expect a SMART system to only show certain Products, right? What is the point of showing me list that has “Pencils, Desk, etc.” when I had older filtered the category to “Groceries”, right?
OK, so the concept of the cascading filters makes sense. We want the system to apply this concept of “showing relevant data” whenever we use a filter. This concept can apply across dimensions. Taking the example above, if I also want to filter on geographies, maybe, I want the system to only show the regions where we have had sales in the Product Category. Right?
This is where your smarts come in. You don’t want the system or technology to apply all sorts of filters. Apply filters across a hierarchy (like Product Category, Product Name) almost always makes sense, but applying this logic across dimensions (when going from Product to Geography) may not always makes sense.
You DON’T want all dimensions to have this concept of “Relevant Values” be applied to it. Because, your system may end up recalculating all the possible values of filters every time you make a click on a filter.
This tip doesn’t always make sense and you will need to find the right balance of applying such logic on certain fields and not all the fields.
Feel free to comment and let me know if you would like additional information.