Collaboration evolution

[jcabot]

I've been thinking about the role I´d like the second panel (the one now called "written pages evolution") to be.

I think that while the first one focuses on overall results achieved by the author (in terms of number of publications), the second one could focus on how the authors has been collaborating to achieve those results.

Therefore, I propose to do the following changes:

• "Activity panel": as it is now or, potentially, adding a line chart on top of the current bar chart with the total page published for the year (not taking into account the number of authors per paper, just total number of pages adding all publications in the year for that author)
• Collaboration panel, showing for every year:

Average number of co-authors for the papers published that year Number of papers for the year when taking into account the number of authors per paper. This would be similar to the idea of average number of pages but I think it's easier to understand. The idea is that I´ve published in one year 3 papers, with 3 authors each (including myself) in the first panel I´d show 3 as a result but on the collaboration panel I´d show 1 (1/3 + 1/3 + 1/3)

Both collaboration metrics are important.

[valeriocos]

the metric "average number of co-authors per year is clear to me. On the other hand, I'd say that the second metric (let's call it "participation") should be read together with the total number of co-authors per year. In particular, if the outcome of the participation metric is low, we don't know whether the author collaborated on a single paper with few co-authors (1/2) or he participated in more papers with many co-authors (1/6 + 1/6 + 1/6). A similar consideration can be done when the participation outcome is high.

In any case, I have just pushed an sql query that contains the average number of co-authors, the sum of co-authors (with duplicates) and the participation value per year for a given author. We can then decide which information we want to show in the collaboration panel.

[jlcanovas]

The metric "average number of co-authors per year" is also clear to me. The second (i.e., "number of papers for the year...") is more difficult to see because the numeric result may not give a clue about how the collaboration was during the year.

What about weighing the result according to some factor (e.g., number of papers considering that all of them were co-authored with the max number of co-authors)? Well, this may sound even more complicated...

Other option could be to indicate the min and max values (e.g., papers with the least and most co-authored values).

[jlcanovas]

Also, for the Activity panel, I'd remove the pages written dimension, I doubt it may reveal something (its behavior mimics the visualization based on papers)

[jlcanovas]

This is what we have now for activity:

Visually attractive but I think that the pink line (number of pages) is not very useful as it behaves as the number of publications.

[jcabot]

It´s normal that it correlates but still I think it's curious to see how many papers of new material you've produced every year. I´d leave it

[jcabot]

"the metric "average number of co-authors per year is clear to me. On the other hand, I'd say that the second metric (let's call it "participation") should be read together with the total number of co-authors per year. In particular, if the outcome of the participation metric is low, we don't know whether the author collaborated on a single paper with few co-authors (1/2) or he participated in more papers with many co-authors (1/6 + 1/6 + 1/6). A similar consideration can be done when the participation outcome is high."

Correct, that's why I proposed both metrics. They go together. We could add in the main panel a metric called "average participation" to complement this as well (same as we have average number of publications per year).

I´d go ahead with this proposal and at least see what kind of results we get. I can´t think of a simpler metric and I don´t want to have the collaboration panel with only one metric (the co-authors one)

[jlcanovas]

Regarding the Activity panel, we'll keep it as it is (with number of pages per year).

Regarding the second panel, the collaboration evolution, it's done showing the two metrics. This is the what we have now: