[Bioc-devel] Package download statistics
richard.pearson at postgrad.manchester.ac.uk
Wed Jun 27 17:14:14 CEST 2007
Thanks for the link - very useful!
I think developers would like to know:
1. How many times their package(s) has(have) been downloaded, ideally
broken down by:
a. Total downloads/number of unique visitors
b. Source/windows binary/OSX binary
c. R download (e.g. using biocLite)/direct download from web page
e. version number
2. How many times each of their package vignettes has been downloaded
3. How many times the package page
I think users would probably like this same information, but for all
packages as a ranked list, and probably also for subsets of packages as
defined by the Bioconductor Subviews.
Having had a look at the url you posted, I think most of this
information is already available, if a little difficult to find, from
the "Pages-URL" section. What I'm not quite sure about is whether this
would include packages downloaded from R (e.g. using biocLite,
install.packages, etc.) - maybe someone could clarify this?
As a first step, I would suggest listing on each package page the number
of unique visitor downloads for the last month and last year of each of
I think www.bioconductor.org might get counted separately from
bioconductor.org, so I'd suggest just adding the two numbers for each of
As a second step I would suggest giving users the option of ranking the
list of packages in each subview by the sum of the above three numbers.
Of course, I have no idea how much work might be involved in doing this,
and how much interest there is from others. Anyone else have views?
Robert Gentleman wrote:
> There is some data here:
> Doing something more, on a per package basis for example, would
> require some more work and before embarking on that, we would like to
> be confident that our developers and users would find such data
> useful. Some suggestions on what would be helpful, would be of
> particular use.
> best wishes
> Richard Pearson wrote:
>> Is there any way of accessing download statistics for Bioconductor
>> packages? If not, would it be possible to make such statistics
>> Bioc-devel at stat.math.ethz.ch mailing list
More information about the Bioc-devel