Vizible Difference

Anywhere on the dashboard, I mean. This is a sequel to Creating Visual Tooltips.

In that post, we mentioned some limitations of the method:

- Data marks, when under a tooltip sheet, are insensitive to mouse hovering

- Transitional lighting effect before a tooltip shows

Here we present a solution by which those limitations can be minimized or removed. The new solution will use a method based on vertical containers, in which, a sheet will collapse if not selected.

The main steps are as follows, assuming you have all the tooltip sheets ready.

1. Create your primary worksheet and drag it to a dashboard canvas. It could be a map or a chart where you want to add some visual tooltips.

2. Drag as many containers (in Floating mode) as tooltips into the dashboard canvas.

[Update: It is reported the tooltips are not working in 10.4. I tested it in 10.5 and confirmed that it is not working either. Will update if there is a workaround.

I wrote a request for bringing the feature back to new Tableau. Please vote it up here

https://community.tableau.com/ideas/9126

]

[Joe Oppelt has figured this out and mentioned this in a forum discussion. At the time, I just didn't understand what he was talking about. I even suggested him to write a post.

Box Plot is a highly effective tool for data analysis invented by the great statistician Professor John Tukey. It gives us the following things:

- Partition of data by quartile.

- Visual spread of each quartile

- Descriptive statistics: max, min, median, upper & lower quartiles.

Quartile is a higher level of details that allows us to understand data at a summary or aggregation level. What made box plot so popular is its simplicity.

[Update: A new method has been posted

http://vizdiff.blogspot.com/2016/10/counting-customers-who-bought-both-and.html]

In marketing nomenclature, this is called market basket analysis. The analysis is aimed at understanding customer purchasing behaviors and the correlations between products.

In Tableau Knowledge Base, there is an article about the topic.

Create Views for Market Basket Analysis

The suggested approach is good for relative small data set.

There came a question in Tableau forum a few days ago:

How to count current (or active) customers who ordered and to whom we delivered goods within last 4 weeks? A corollary question: what are the count of active customers at a past date?

This belongs to a general class of similar problems, which are all about counting.

This is a sequel to Taking Stock with Start and End Dates. We will present a new way of taking stock. It will make the calculation much easier.

Counting active orders is equivalent to taking stock. An order is active before it is shipped. So an order is like part of the inventory before it leaves the warehouse.

In the previous post, we presented 2 approaches to solve the problem.

[Update] Found a new way of doing the same. It is by pivoting. Check out my video.

[Measure Names] is sitting in the Dimensions shelf in Tableau. But it is not a true dimension, because we can't do table calculations along it for example. So we have to resort to some unusual technique to make it a dimension.

Someone got a single row of data with multiple columns of which most are measures. Those are the result of the upstream filtering and aggregation.

This post follows a few others on the same topic:

Infectious Diseases in California 2001-2014: Trends

Infectious Diseases in California 2001-2014: Top Diseases and Counties

Infectious Diseases in California 2001-2014: Men vs Women

(Click image to enable interactivity)

Box plot has been an effective way to give us quick insights into the distribution of data. It is done by partitioning the data by quartile.

This is part of my continuous experiment to better visualize spatial data. An earlier of post is here:

Spatial Analysis by Quartile Partitioning

It was a revelation to me when I saw the California population distribution like above. Half of the Californians are squeezed in between Los Angeles and San Diego.

When I was working on visualizing the disease cases and ratios in California, I wanted to see the spatial distribution of the disease cases.