Module 3 Exercise 3: Assessing collaboration

(Chad Purdy) #41

I chose Sway, an i3 compatible Wayland compositor. The first image below, shows that the activity over the course of the development has been relatively active, but has really picked up again this year.

The second image, shows that the number of merged pull requests and closed issues in the past week is pretty high.

This type of information would give some insight into whether groups used the designated time for their project wisely, or if they waited until the last minute to start working. In addition, I can see top contributors, which would help with individual assessment.

(Max Hudnell) #42

The open source project I chose to look at is keras. My two pieces of evidence that it is an active open source project is that it has 32668 stars, and the code frequency graph shows that there have been constant commits over the past year.

These tools help assess collaboration with group work, because it enables you to see when development happens, and who worked on the project.

(Rishabh Bansal) #43

HI @mozzadrella, this my module 3 exercise 3.4 submission. :slight_smile:

I choose Add-ons project repository by Mozilla! It shows 48 authors and 36 merged PR’s in last week, quite a active project? also, 64 active issues, with 50 closed issues shows great engangement!

(Chris Cannon) #44

Great responses @theportablegeek! I appreciate the screenshots of each item you’re addressing. Hopefully these leads to full evaluations for your students!

(Chris Cannon) #45

Good point, @ConsoleFriend :joy: I’m hoping to roll out automated testing this semester so that I can quickly see whether or not student’s code is functional and move on to helping them improve their coding process and collaboration skills.

(Chris Cannon) #46

Agreed @rmarku! :100:

(Chris Cannon) #47

How would these tools be helpful as an instructor evaluating students work? :thinking:

(Chris Cannon) #48

Nice @rishabh-bansal! How can these same metrics help you evaluate your student’s participation in assignments?

(Eduardo Gutierrez) #49

Exercise: Assess a live project

  1. Find an active open source project.

HardInfo is a system profiler and benchmark for Linux systems. It is able to obtain information from both hardware and basic software, and organize them in a simple to use GUI.

Its an active project but with some ups and downs…

  1. Point to 2 pieces of evidence that the project is active.

Regular commits made this year, in june

It has as said before some ups and downs, and seems that after some tiem inactive this 2018 has turn back to live…As you can watch in contributors seems that original dev pereira has left the developement of the aplication to bp0

3. These pieces of evidence can be:

  • Number of forks or stars: 43 forks and 136 star

Reflect on how you might use these insights to assess collaboration in group work:

Well i dont plan to use group work often. Just one assignement just to know how it works. The most interesting thing could be the contributors ( in insights) where we can watch the activity of the different members of the group, if there is someone that has worked a lot or if someone has not worked at all.

Also would be interesting to watch commits, who and when made them, if they have been working in a progressive way, or only worked the last day or so.

And the third important element would be the code frequency. Its a pity that you cant configure different variables like y-axis values, the period of time analyzed ( instead per weeks, per month or define the begin and end date of frequency)

(Chris Cannon) #50

How can these same types of states help you evaluate student work @ygorcanalli?

(Ygor Canalli) #51

This can help me evaluating students by seeing their commitment, working frequence and overall contribution.

(Chris Cannon) #52

Grading participation is always difficult, especially in computer science where we tend to set very clear, measurable criteria. One method I would recommend is publishing a rubric with both objective and subjective criteria. For example, 80% of the students grade could come from their code meeting your guidelines, 10% could come from code being more or less efficient, and 10% could come from participation.

I would start with student’s receiving full credit for participation and they can lose points if it seems like they did not make a significant contribution.

(Chris Cannon) #53

Exactly! Its hard not to laugh when a student will tell me they need an extension on an assignment but I can see they started it the day before it was due :joy:

(Miren Berasategi) #54

Yes, that is my idea as well. But I find my students doing “bulk” commits most of the time, that is, working somewhere “outside” GitHub/their text editor, then copying and pasting everything into one single commit – which makes it all pointless. I seem unable to make it clear that ALL their work needs to be recorded into GitHub commits! Maybe I just need to award direct points to (or substract them for the lack of) regular commits… :sweat:

(Miren Berasategi) #55

I could possibly design a metric that would award points for

  1. regular contributions (not tending to peak too much toward the deadline
  2. number of commits. Which I happen to not like too much (or find myself unable to design in a way that is representative) because, as I said in the previous comment, my students commit very irregularly (some commit every typo they correct, other commit all the submission at once)


(Chris Cannon) #56

I think it’s worth playing with, enforcing the process is difficult but ultimately helps your students get the most out of the tool! Horror stories of broken computers and lost data can only go so far to convince students :joy:

(Docmilo) #57

I choose CppCoreGuidelines:

1,608 commits, 8 branches, 206 contributors

(Chris Cannon) #58

@Docmilo what do these statistics tell you about the contributions? How can these same metrics help while evaluating student work?

(Docmilo) #59

Hey @ccannon94 Thanks for the comment. In the past I have used metrics from GitHub to determine student engagement in assignments. Ideally I want to see a sustained level of engagement (measured through commits) from the assignment being issued up to and including the deadline. Plus I want to see commits over the course of the assignment and not a panicked rush at the end. GitHub stats can also identify who make the most / significant contributions to the work.

(Chris Cannon) #60

I’m never sure whether to laugh or cringe when I receive a notification the night before a multi-week assignment is due that a student just started :sweat_smile: