Comments on: Antibiotic Effectiveness – A Study of Chart Types

By: Anonymous

Anonymous — Sat, 19 Dec 2009 17:41:54 +0000

My ignorance is showing, but can the author or anyone cognizant explain the use of the mic’s in culture and sensitivity reports at the hospital in terms of choosing one antibiotic over another.
For instance: the S means sensitivity, as opposed to R for resistant, but the S, or SS might have <.5 S and another antibiotic might have <4 S.

What is the difference say between two sensitive antibiotics where one has <.5 or 1 or 2, versus another that has
<4, or 5 etc etc in terms of efficacy etc. What does this numerical difference above mean in terms of efficacy or complications etc.

Thank You

Mark

By: Jon Peltier

Jon Peltier — Sun, 25 Jan 2009 00:24:20 +0000

I agree that in general it’s best to avoid the connecting lines, but the alternatives are less effective. Markers without the lines are more difficult to visualize, and I abandoned the various bar chart variations before moving to the line and marker approaches.

By: Yawar Amin

Yawar Amin — Sat, 24 Jan 2009 23:04:51 +0000

I’m still not convinced that connecting lines between data points in categorical data is a good idea … just feels inherently wrong. Still, I’ll admit Xan Gregg’s point that it does make it easier to see obscured data points.

By: Xan Gregg

Xan Gregg — Sun, 11 Jan 2009 19:00:30 +0000

Nice work. I usually try to avoid connecting points on categorical scales, but your plot shows another use for such lines. It allows the reader to see where the obscured data points are.

By: Burtin Antibiotic Illustrations « FORTH GO

Burtin Antibiotic Illustrations « FORTH GO — Sun, 11 Jan 2009 18:51:45 +0000

[...] Peltier has a write-up of his contest entry. It has all the data of Burtin’s original in a much better rectangular [...]

By: Jon Peltier

Jon Peltier — Sat, 10 Jan 2009 13:21:52 +0000

Jaemyung -

Thank you for your comments. I have actually incorporated a couple sentences to help clarify things.

I think your question about sorting was really about how the data is sorted. There is no capability to sort data within the chart, but you can get the sorting you want in an existing chart by sorting the data upon which the chart is based. Since this specific article was less about the mechanics of making these charts, and more about how I approached the objectives of the contest, I left out a lot of the mechanical details. Other articles I write show these mechanics in much greater detail.

By: JaemyungKim

JaemyungKim — Sat, 10 Jan 2009 05:19:41 +0000

Dear Dr. Peltier,
I am just a junior student from some middle states. These illustrations are awesome. I will try this contest, but i don’t think I will make this clearly and intuitively.

I wanted to comment, right after your first chart, you added “The larger the concentration, the less effective an antibiotic is at inhibiting bacteria growth, so the longer bars above indicate lower effectiveness, which is counterintuitive.” and then, you reverse the direction of log scale. I hope you also commented that the longer bar would mean the better antibiotic with higher effectiveness.

and, that the chart would be easier, for beginner like me, if you added the description of “sort by” as part of the bar charts, on the 4,5, and 6th charts.

I haven’t learned many statistics, Excel, and graphic programs. Still you work look aesthetic for me. Hope you have a great result from the contest.

Jaemyung Kim

By: Jon Peltier’s Excel entry to the Chance Magazine contest « visualization, etc.

Thu, 08 Jan 2009 23:10:08 +0000

[...] showing some pretty effective plots (and he also hates pie charts – awesome) on his blog, and his entry to the Chance Magazine’s contest (contest mentioned here) is no [...]

By: Jon Peltier

Jon Peltier — Wed, 07 Jan 2009 12:00:51 +0000

SB -

I’ve thought about how I might accomplish this highlighting. I suppose it would be easiest to format the text differently, i.e., using bold or italic styling or a different font color. However, this decouples the staining from the antibiotic behavior, and I think it would be harder to get the fact that the Gram-positive bacteria are also the one that penicillin is most effective against. This kind of relationship is what leads to insights about mechanisms of drug action. For example (and I’m speculating wildly, because I’m not a biochemist), perhaps penicillin is absorbed into the cell through a mechanism like that which absorbs the Gram stain.

By: SB

SB — Tue, 06 Jan 2009 20:09:50 +0000

I think highlighting the bacteria name instead of the grid in the chart might be cleaner.