Wave 103: Exploring an Airline Dataset

First thing to do is to understand the grain level of the dataset, that is what does a row in the dataset indicate.  

How many flights are there from SFO to ORD and vice versa. 

Query: Measures, dimensions, filters (3 parts). 

The  Filters : ORD and SFO as origin and destination and vice versa.  

To break into individual buckets add a grouping. Add Origin.  

Destination grouping not necessary as its filtered. However, adding this will help the end user. 

Generally we keep the grain level consistent in a Dataset. 

Show % cancelled flights by destination for United flights.  

Filter: United, Origin, Cancelled. Group by Destination. Measure % of flights.  

Measure % can be done by Stacked bar and Group by Destination.  

 Normalize function is only available for a Stacked Bar Graph. Not for anything else.  

When values in groupings are given they are OR'd and between filters are AND'd. 

You can show only what bars you want to with the FOCUS option. This is done by highlighting the bars you need and going to settings and choose 'Focus' 

If you want non- contiguous rows, you can just select a bar on it own . 

What is Cartesian Explosion?  

It is the same as Cartesian Cross Product.  

Don’t use this ever for the end user as it is very difficult to consume for the end user.