Enhanced tables, and supercharged spreadsheets with in-cell tech

Old-timer Chris P. sent me to this Bloomberg article about Vanguard ETFs and low-cost funds (link). The article itself is interesting, and I will discuss it on the sister blog some time in the future.

Chris is impressed with this table included with the article:

This table indeed presents the insight clearly. Those fund sectors in which Vanguard does not compete have much higher costs than the fund sectors in which Vanguard is a player. The author calls this the "Vanguard effect."

This is a case where finding a visual design to beat this table is hard.

For a certain type of audience, namely financial, the spreadsheet is like rice or pasta; you simply can't live without it. The Bloomberg spreadsheet does one better: the bands of blue contrast with the white cells, which neatly divides those funds into two groups.

If you use spreadsheets a lot, you should definitely look into in-cell charts. Perhaps Tufte's sparkline is the most famous but use your imagination. I also wish vendors would support in-cell charts more eagerly.

Here is a vision of what in-cell technology can do with the above spreadsheet. (The chart is generated in R.)

 

 

 

Comments

[ZBicyclist]

Yes, the table works nicely. I've grown to distrust any graphic put out by the financial services industry as likely to be misleading in some way, so I always want to see the actual numbers.

In terms of the article, John Bogle is a giant. They should give him the Nobel Prize for Economics for all the money he has saved investors (or maybe Vanguard as a whole, the way they gave MSF -- Doctors without Borders -- the Nobel Peace Prize). His investment advice is very sound for the average person who doesn't have the time or inclination to micromanage.

[Shampshire]

I guess it depends on the purpose. The table is best if we want to look up exact numbers, but I'm not sure it's the clearest way to show the pattern.

I completely agree with you about in-cell charting, but would there be anything wrong with a scatter plot treatment?

Super-quick sketch here

[Mark Gandy]

Shampshire, looks great and also tells the/a story quickly. But also like just the table and Kaiser's visual.

[J Santiago]

Just found out your blog and started learning R for my PhD. I was wondering how you did those plots -- ggplot?

[Kaiser]

Shampshire: Thanks for the alternative. In this case, there are many ways to make it work.

J Santiago: R is one of those languages that allow you to construct a graph by drawing items piece by piece. Think of each item on the chart as existing in a "cell" (within a spreadsheet) so the basic code places the row label, then the square symbol, then the line segment, and finally the dot. You can wrap that in a function. Then you run the function row by row. Finally you apply the title and column headers.

The only other trick is to adjust the settings of the spacing around the chart so you can accommodate the text.

Feel free to email me and let me know which part you are having trouble with.

[J Santiago]

I understand the concept, but since I'm quite new to R (learning it since a month!) I have no idea where to start. I've been using ggplot for my graphs, but you seem to have "hand built" yours. Could you point to a good resource on how to learn to do that?