Stephen Royle

**Note**: this is not a serious blog post.

Neil Hall's think piece in [Genome Biology on the Kardashian
index](http://genomebiology.com/2014/15/7/424) (K-index) caused an
online storm recently, spawning
[hashtags](http://jasonya.com/wp/alternatesciencemetrics-that-might-actually-work/)
and
[outrage](http://www.redink.cc/2014/08/03/let-me-fix-that-for-you-neil-hall/)
in not-so-[equal measure](http://kateclancy.com/jokes-that-dont-work/).
Despite all the vitriol that headed Neil's way, very little of it
concerned his use of Microsoft Excel to make his plot of Twitter
followers vs total citations! Looking at the plot with the ellipse
around a bunch of the points and also at the equations, I thought it
might be worth double-checking Neil's calculations.

In case you don't know what this is about: the K-index is the ratio of
actual Twitter followers (\\F\_{a}\\) to the number of Twitter followers
you are predicted to have (\\F\_{c}\\) based on the total number of
citations to your papers (\\C\\) from the equation:

\\F\_{c}=43.3C\^{0.32} \\

So the K-index is:

\\K-index=\frac{F\_{a}}{F\_{c}}\\

He argues that if a scientist has a K-index \>5 then they are more
famous for their twitterings than for their science. This was the most
controversial aspect of the piece. It wasn't clear whether he meant that
highly cited scientists should get tweeting or that top-tweeters should
try to generate some more citations (not as easy as it sounds). The
equation for \\F\_{c}\\ was a bit suspect, derived from some kind of fit
through some of the points. Anyway, it seemed to me that the ellipse
containing the Kardashians didn't look right.

I generated the data for \\F\_{c}\\ and for a line to show the
threshold at which one becomes a Kardashian (k) in IgorPro as follows:

`Make /o /N=100000 fc`  
`fc =43.3*(x^0.32)`  
`Duplicate fc k //yes, this does look rude`  
`k *=5`  
`display fc, k //and again!`

[![K-index](https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?resize=300%2C187&ssl=1){.alignleft
.size-medium .wp-image-175 loading="lazy" decoding="async"
attachment-id="175" permalink="https://quantixed.org/k-index/"
orig-file="https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?fit=2163%2C1354&ssl=1"
orig-size="2163,1354" comments-opened="1"
image-meta="{\"aperture\":\"0\",\"credit\":\"\",\"camera\":\"\",\"caption\":\"\",\"created_timestamp\":\"0\",\"copyright\":\"\",\"focal_length\":\"0\",\"iso\":\"0\",\"shutter_speed\":\"0\",\"title\":\"\",\"orientation\":\"0\"}"
image-title="K-index" image-description="" image-caption=""
medium-file="https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?fit=300%2C188&ssl=1"
large-file="https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?fit=640%2C401&ssl=1"
width="300" height="187"
srcset="https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?w=2163&ssl=1 2163w, https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?resize=300%2C188&ssl=1 300w, https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?resize=768%2C481&ssl=1 768w, https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?resize=1024%2C641&ssl=1 1024w, https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?w=1280&ssl=1 1280w, https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?w=1920&ssl=1 1920w"
sizes="(max-width: 300px) 100vw, 300px"
recalc-dims="1"}](https://i0.wp.com/quantixed.org/wp-content/uploads/2014/08/k-index.png?ssl=1)

This plot could be resized and overlaid on Neil's Excel chart from
Genome Biology. I kept the points but deleted the rest and then made
this graph.

The Kardashians are in the peach zone. You'll notice one poor chap is
classed as a Kardashian by Neil, yet he is innocent! Clearly below the
line, i.e. K-index \<5.

Two confessions:

1.  My K-index today is 1.97 according to Twitter and Google Scholar.
2.  Embarrassingly, I didn't know of the business person who gave her
    name to the K-index was until reading Neil's article and the ensuing
    discussion. So I did learn something from this!

---

The post title is taken from "Vitamin K" by Gruff Rhys from the Hotel
Shampoo album.


<strong>
 Note
</strong>
: this is not a serious blog post. Neil Hall’s think piece in Genome Biology on the Kardashian index (K-index) caused an online storm recently, spawning hashtags and outrage in not-so-equal measure. Despite all the vitriol that headed Neil’s way, very little of it concerned his use of Microsoft Excel to make his plot of Twitter followers vs total citations!

Vitamin K

In the lab we use IgorPro from Wavemetrics for analysis. Here is a
useful procedure to plot all XY pairs in an experiment. I was plotting
out some cell tracking data with a colleague and I knew that I had this
useful function buried in an experiment somewhere. I eventually found it
and thought I'd post it here. I'll add it to the [code section of the
website](http://mechanochemistry.org/royle/papers/#tab=code){target="_blank"}
soon. Looking at it, it doesn't look like it was written by me. A search
of IgorExchange didn't reveal its author, so maybe it was me. Apologies
if it wasn't.

The point is: if you have a bunch of XY pairs and you just want to plot
all of them in one window to look at them. If they are 2D waves or a
small number of 1D waves, this is straightforward. If you have hundreds,
you need a function!

An example would be fluorescence recordings versus time (where each time
wave is unique to the fluorescence trace) or XY co-ordinates of a
particle in space.

To use this procedure, you need an experiment with a logical naming
system for 1D waves. something like X_ctrl1, X_ctrl2, X_ctrl3 etc. and
Y_ctrl1, Y_ctrl2, Y_ctrl3 etc. Paste the following into the Procedure
Window (command+m).

    Function PlotAllWaves(theYList,theXlist)
        String theYList
        String theXList
        display
        Variable i=0
        string aWaveName = ""
        string bWaveName = ""
        do
            aWaveName = StringFromList(i, theYList)
            bWavename = StringFromList(i, theXList)
            WAVE/Z aWave = $aWaveName
            WAVE/Z bWave = $bWaveName
            if (!WaveExists(aWave))
                break
            endif
            appendtograph aWave vs bWave
            i += 1
        while(1)
    End

After compiling you can call the function by typing in the Command
Window:

    PlotAllWaves(wavelist("x_*", ";", ""),wavelist("y_*", ";", ""))

You'll need to change this for whatever convention you are using for
your wave naming system. You will know how to do this if you have got
this far!

This function is very useful for just eyeballing the data after you have
imported it. The databrowser shows only one wave at a time, but it is
preferable to look at all the waves to find errors, spot outliers or
trends etc.

Edit 28/4/15: the logical naming system and the order in which the waves
were added to the experiment are crucial for this to work. We're now
using two different versions of this code that either a) check that the
waves are compatible or b) concatenate the waves into a 2D wave before
plotting. This reduces errors in plotting.

---

The post title is taken from All Together Now -- The Beatles from the
Yellow Submarine soundtrack.


In the lab we use IgorPro from Wavemetrics for analysis. Here is a useful procedure to plot all XY pairs in an experiment. I was plotting out some cell tracking data with a colleague and I knew that I had this useful function buried in an experiment somewhere. I eventually found it and thought I’d post it here.

Messaggi di Rogue Scholar

Vitamin K

All Together Now