Activity Index

From SONIVIS:Wiki

Jump to: navigation, search
Activity Index

This metric is not implemented in any version. You can report this as a Bug.


This metric attributes to an actor.


This metric can be calculated for the connector MediaWiki.



Description

Average amount of added content in kByte

Explanation

Shows the average amount of added content of an author.

Basic assumptions

Maybe this metric can also be called Externalization Level Index (German:Externalisierungsniveau) because the the "quality" and the amount of a contribution is measured. Creating content is presumably an action of the author for externalizing knowledge.

In a wiki there is a variety of possibilities for an author to take part and measure her/his created content. We could measure the number of pages s/he worked on or the rate of made changes. Looking at both options of measuring her/his participation we could use the average to calculate an average rate of made changes on pages. The considered pages for this calculation are only pages this author worked on.

Using this measurement we only get a quantitative issue of the changes per page of an author. We can not say anything about the quality of the made changes. We do not consider that the contributor just added a comma or wrote a whole new chapter. If we want to include one aspect made for quality we should consider at least the amount of contribution.

For the activity calculation we should not only include the revision count but also the distribution of edits over pages and the amount of contribution.

Calculation

To show the activity of an author we do not need a network for the calculation.

involved measures:

  1. first considered calculation: AN(n_i)=\frac{EC(n_i)}{AC(n_i)}+\frac{AmC(n_i)}{EC(n_i)}
    • there is no evaluation of the cohesion between articles, contributions and the amount of contributions
    • we do not get a conclusion about the expressiveness of all measures positioned in cohesion
  2. second considered calculation: AN(n_i)=\frac{EC(n_i)}{AC(n_i)*AmC(n_i)}
    • this calculation is wrong for our case: authors with a less number of edits with a big amount gets a smaller AN than an author with a high number of changed article each with a small-sized amount of contribution
    • all authors with s smaller number of edits but each with a big amount of contribution should get a greater AN than authors with much minor edits
  3. third considered calculation: AN(n_i)=\frac{x*AmC(n_i)*y*AmC(n_i)}{AC(n_i)}
    • seems to fit, but please look at the example for correcting this statement
    • This example shows us that changing x or y that not influence the outcomes. The only chance to consider these variables are with the sum of AmC and EC
  4. current possible calculation: A_{I}(n_i)= \frac{AmC(n_i)}{\frac{|EC(n_i)|}{|AC(n_i)|}}
    • What tell us small values? What tells us a high value?

Example

Ai AC(Ai) EC(Ai) AmC(Ai) AN(Ai) = (x * AmC(Ai) * y * EC(Ai)) / AC(Ai)

(x = 0,6; y = 0,4)

AN(Ai) = (x * AmC(Ai) * y * EC(Ai)) / AC(Ai)

(x = 0,6; y = 0,4)

AN(Ai) = (x * AmC(Ai) + y * EC(Ai)) / AC(Ai)

(x = 0,6; y = 0,4)

A1 1 12 20Kb ((0.6*20)*(0.4*12))/1 = 57.6 ((0.4*20)*(0.6*12))/1 = 57.6 ((0.6*20)+(0.4*12))/1 = 16.8
A2 2 24 20Kb ((0.6*20)*(0.4*24))/2 = 57.6 ((0.4*20)*(0.6*24))/2 = 57.6 ((0.6*20)+(0.4*24))/2 = 10.8
A3 2 24 40Kb ((0.6*40)*(0.4*24))/2 = 115.2 ((0.4*40)*(0.6*24))/2 = 115.2 ((0.6*40)+(0.4*24))/2 = 16.8
A4 20 30 1Kb ((0.6*1)*(0.4*30))/20 = 0.36 ((0.4*1)*(0.6*30))/20 = 0.36 ((0.6*1)+(0.4*30))/20 = 0.63
A5 1 30 12Kb ((0.6*12)*(0.4*30))/1 = 86.4 ((0.4*12)*(0.6*30))/1 = 86.4 ((0.6*12)+(0.4*30))/1 = 19.2
A6 3 3 4Kb ((0.6*4)*(0.4*3))/3 = 0.96 ((0.6*4)*(0.4*3))/3 = 0.96 ((0.6*4)+(0.4*3))/3 = 1.2
A7 12 40 5Kb ((0.6*5)*(0.4*40))/12 = 4 ((0.4*5)*(0.6*40))/12 = 4 ((0.6*5)+(0.4*40))/12 = 1.583333
A8 12 40 10Kb ((0.6*10)*(0.4*40))/12 = 8 ((0.4*10)*(0.6*40))/12 = 8 ((0.4*10)+(0.6*40))/12 = 2.333333

This example shows us that changing x or y that not influence the outcomes. The only chance to consider these variables are with the sum of AmC and EC. We have to think once more about a calculation which makes sense!!!

Open Questions

  • Should minor edits be involved?
  • There is no influence of time in this calculation! Over which time slot was the author active?
  • Does an author have a special contribution time? Does s/he only get active in special times?
  • Should the Edit Count more weighted as the Amount Contribution?
  • Does a generalized calculation for the entire wiki make sense?
    • draft: AN(wiki)=\frac{\sum_{i=1}^{g}AN(n_i)}{g}
  • How should we normalize? Does a normalization make sense?
  • For correct calculation in a network we should include only pages which are part of the network?!
  • Should we only consider articles?
  • Does the Amount Contribution include all changes or just one?
  • include rigor & diversity
Retrieved from "http://www.sonivis.org/wiki/index.php/Activity_Index"
Personal tools