APOD web site statistics

[Joan Girones]

Statistics about

- APOD reader’s behavior

- analysis systems (Google Analytics, Alexa, server tracking…)

- curiosities

And any topic related with APOD statistics.

If you are an APOD mirror site operator and you want to share your data, that’s the place.

(Some data are only based on Catalan APOD mirror, right now from January 1st 2002; if so, it will be warned)

[Joan Girones]

Within the global trend in which the users don’t click the links, there is a special kind. I name it “abstracts”, despite these links are not abstracts in every instance.
Some examples are:

http://adsabs.harvard.edu/abs/1972ApJ...178..623T
http://adsabs.harvard.edu/cgi-bin/nph-b ... .135..181M
http://arxiv.org/abs/1003.4860

The average reading time on this kind of pages is roughly:

5 seconds in the last 30 days.
9 seconds in the last 6 months, with 2 exceptions.
In the last 6 months there have been 60 abstract links.

That means that, in fact, the users of Catalan mirror don’t read these pages.

It’s sure that the language is an important factor but, on the other hand, there are also pages in English which the users spend much more time (like English Wikipedia).

(data from October 12 2010 – Catalan APOD mirror)

[RJN]

That means that, in fact, the users of Catalan mirror don’t read these pages.

First, I welcome discussions of this type as they have the potential to make APOD stronger. Next, I am a bit surprised by these results but not very surprised. A perhaps helpful bit of background is that that the APOD explanation is meant to be a resource on a variety of levels, not just on the most popular level. APOD is read by many professional astronomers and by many people who teach astronomy. These people, although in the vast minority by number, might appreciate a bit more depth to the explanations than the text given and the popular links given. Still, these people, especially the teachers, might be able to translate the additional depth given by the journal links into a better lecture or presentation. In fact, I have had a professional astronomer tell me just that -- that he really doesn't read the APOD text but looks for links that give more detailed information. I also picture, in my mind's eye, that some students do reports on APOD-related images and would also appreciate additional depth. Last, some people might not believe that the information given in the APOD explanation is at the state-of-the-art for what professional astronomers might believe. These journal links might work to correct this potential misunderstanding.

Perhaps your underlying point is that we are saddling the APOD reader with unwanted links and hence watering down the whole "APOD experience". Unfortunately, that is correct. Therefore each link must balance the good and the bad aspects of including it. Therefore, perhaps, there is no right or wrong answer to each question on link inclusion, but rather each link choice is a journalistic style not unlike word choice.

- RJN

[rstevenson]

...
http://adsabs.harvard.edu/abs/1972ApJ...178..623T
http://adsabs.harvard.edu/cgi-bin/nph-b ... .135..181M
http://arxiv.org/abs/1003.4860

The average reading time on this kind of pages is roughly:

5 seconds in the last 30 days.
9 seconds in the last 6 months, with 2 exceptions.
In the last 6 months there have been 60 abstract links.

That means that, in fact, the users of Catalan mirror don’t read these pages.

I wouldn't jump to that conclusion too fast. For example, arxiv.org links take one to the same page layout always, and always there is a link to a PDF of the paper near the top-right corner. I can quickly scan the abstract and then, if I want to, click that PDF link to download the whole paper. I do this in a very short amount of time, often less than 10 seconds.

The only sure way to know what people are doing on web pages is to watch them by looking over their shoulder. Statistics can't tell you the whole story and may in fact be misleading.

Rob

[Joan Girones]

Statistics can't tell you the whole story and may in fact be misleading.

I’m absolutely agreed with the first part of your statement. Furthermore: statistics never tells the whole story.
Buy they only may be misleading if are taken as more than a trend; just a clue to begin understanding a phenomena.

Besides, on the case which I showed:
- I used only an analysis tool: “Google Analytics”; it is a system on client-side. It must also be used a tracking system on server-side to get better results.
- Spain is one of the countries of the European Union with a lower knowledge of English language. That is important on the kind of links about we are talking.
- Finally, the potential readers from Catalan mirror (the sample data) is more little than any other from non English APOD mirrors. A minor sample data generate results less reliable.

But, despite that, and also answering to RJN, I think that statistics are interesting if they are interpreted carefully.

[Joan Girones]

Internal links

The APOD web site internal title more linked (178 times) is “Fractal Interstellar Dust Up-Close”, available on:
http://antwrp.gsfc.nasa.gov/apod/ap030706.html
http://antwrp.gsfc.nasa.gov/apod/ap990509.html
http://antwrp.gsfc.nasa.gov/apod/ap980104.html
http://antwrp.gsfc.nasa.gov/apod/ap961119.html

External links

Particular web sites

Wikipedia has been linked 5,230 times.
The first link to Wikipedia was made on 2003 October 29 – http://antwrp.gsfc.nasa.gov/apod/ap031029.html - on the word “protons”.

YouTube has been linked 517 times.
The first link to YouTube was made on 2007 February 13 - http://antwrp.gsfc.nasa.gov/apod/ap070213.html - on the word “explode”.

Ranking of more linked web sites

The winner is… undoubtedly, the site www.seds.org/ which has been linked 1,103 times.

http://www.astro.wisc.edu/ - 279 times
http://www.hawastsoc.org/ - 166 times
http://coolcosmos.ipac.caltech.edu/ - 153 times


The data are from Catalan APOD mirror, therefore they are roughly.
Dataset: from January 2001 (*) to December 10 2010

(*) Catalan APOD mirror began on May 6th 2002, but there are pages translated back in time until January 1st 2001, as well as some pages which are important.