1. Visualizing all subsets
2. Insight from subset exploration
The result of the last exercise is a plot showing very clearly that the total fare minus tips for credit card transactions has roughly the same distribution as cash transactions. This suggests that the recorded cash transaction amount does not include the tip amount.
We were able to get to this interesting conclusion by zeroing in on one small subset of the data that we could freely explore without worrying about scalable visualization methods.
3. Visualizing all subsets
We have made an interesting discovery in one subset of the data. We may be fine to assume that this phenomenon is the same for all of the other routes in the data.
4. Visualizing all subsets
Suppose we wish to extend this analysis to all of the routes that exist in the taxi dataset to ensure this phenomenon holds, or to potentially discover routes where the observed phenomenon does not hold.
5. Visualizing all subsets with Trelliscope
We have learned that faceting is a great way to make visual comparisons. However, there are over twenty thousand unique routes in this dataset. We can't possibly look at a faceted display with twenty thousand panels!
6. Visualizing all subsets with Trelliscope
Faceting is a simple and very powerful visualization approach, but it begins to be difficult to apply when the dataset gets large and the number of natural subsets of the data to facet on becomes too large to view all at once.
In the following chapter, we will explore an approach to visualize larger datasets in detail through faceting with Trelliscope.
7. See you in Chapter 2!
See you in Chapter 2!