Module 3 Exercise 3: Assessing collaboration

(Alexander L. Hayes) #21
  1. tensorflow/tfjs
  2. The project is fairly new, but lists 8 merged pull requests, 2 closed issues, and 17 new ones in the last week.

Some of these insights could be helpful, but just looking at the metadata (in these tabs) may not tell the entire story of an open source project.

For example: the “Contributors” tab under Insights is ordered by the number of commits a person made. This is not necessarily the same as “the person who did the most work” though it may be correlated.

(Vicente Cubells) #22

Module 3 Exercise 3

I chosen Minikube, a very interesting active project with 2,589 commits, 37 releases, > 200 collaborators, > 8000 stars, > 1000 forks, etc.

As a teacher, this metrics allow me to follow each student’s work an how each of them are contributing to the group assignments.

(Ryan Schuetzler) #23

I chose to look at the Django project:

  2. To assess activity, I looked the recent contributors graph, and the list of recently merged pull requests. Both have lots of activity in the past couple days (or even hours). These could be useful for assessing student work throughout a semester project, making sure things are moving along the whole time rather than a big cram at the end of the semester.

(Danny de Vries) #24

A friend of mine is working on a headless CMS. The core repository has gained alot of stars and the maintainer goes to through the issues every once in a while.

Most of the team is actively working on the API component.

(Joachim Francois) #25

I took the Blazor github page for this exercise.

When looking at the Contributors there’s quite a few actively contributing to the project.

The actual commits are frequent since the start of the project.

The ratio of Merged Pull requests and Closed Issues in the past month are also indicating a good activity.

When it comes down to potential assessment of collaboration in group work I’ll be looking for a couple of things. First of all if the work has been somewhat equally devided, not only taking the amount of code that has been commited into account but also the kind of code they worked on.

Preferably the team has been using branches in a reasonable way while working on the group project and I can find a trace of issues they came across in the Issues list that’s available via github.

Those on top of functional code would be a rather great assessment for the team :wink:

(Rohith Pudari) #26

(Mark Patterson) #27

I chose Sonic Pi because I like the idea of using music and sound to teach coding.

The project is maybe not as active as some of the other ones mentioned in this thread, with only one merged pull request in the last month. The commit rate also isn’t as high as it has been in the past. This would suggest that the project is feature complete (or at least, no new features are currently being worked on), and that it’s mostly just bug fixes for now. It is also obvious that most of the work has been done by one person.

For assessing collaboration, the commit metrics are going to be useful (I’ll be curious to see who’s doing work outside of class (and at what time of day/night;)) The contributors page, as well as the number of commits, it also shows the number if insertions and deletions, which gives a better idea of the amount of work each student is doing. But like what others have said, you still have to drill down and look at the actual code that is being contributed.

I like what @ConsoleFriend said about also looking at branch usage and hopefully the issue queue.

I also like @mikecrabb 's idea about getting the students to reflect and explain why the contributions page looks the way it does at the end of the project.

(Roberto Martínez Román) #28

Code frequency let me see if students are working or not.

(Dr. Ayaz H. Khan) #29

Project: facebook/react

  • Recently Closed issues highlight that the project team is actively handling issues right now.

  • Sufficient number of commits within a week time.

(Cwoods88keys) #30

I choose GnuCash. They had a new release come out this month and was wondering if they were steadily working on it over time - or were they “cramming” at the last minute the way many of my student do. GitHub showed commit history going back over 20 years! That historical data did make it harder to see who was currently active. It’s nice to see the list sorted by the number of commits/additions/deletions and, in a short semester, I would not have the top contributors dating back to 2000.

(Mrculp Mhs) #31

Recent commits

Recent milestones. One 97% complete and other 11%. Also 34 closed milestones

Viewing the commits, milestones, insights, etc will help keep updated on students progress. Often a student acts like they are being productive, but fail to produce work. Looking at the data in the repository will give me an accurate view of work being produced and by whom.

(Theportablegeek) #32

@mozzadrella The active open source project that I was able to locate was youtube-dl, an open source project from developer Ricardo Garcia. This project was found via a review of weekly trending repositories through GitHub Explore, its presence no doubt influenced by a massive star rating and number of forks (see below)

Live since 2008, activity has stayed consistent since it’s peak in mid-2015 with more than 16,000 lifetime commits and 622 lifetime contributors. (see below)

While lifetime commits and contributors are of interest, they do not demonstrate current activity; for that we can instead look to recent insights. Over the past year youtube-dl has seen an average of approximately 20 commits per week, with many weeks well exceeding that average (see below)

Looking at the past month, we can see that this strong project activity has continued through the many active and closed issues, as well as nearly 100 commits from 8 different authors.

The insights provided through GitHub are incredibly valuable to assessment of group work, providing a quick-and-easy snapshot of not only who is committing but when, what and how often. This data takes so much of the guess work out of the evaluation process, allowing educators to not only make more valid and reliable assessments but to provide identify (and provide support to) students who may need a push in the right direction.

(Ygor Canalli) #33

I choose the open source project PyTorch, a Tensor Computation and Deep Learning framework with strong GPU acceleration.

Looking to commits, we see that there is a lot of recent commits. Looking to last 7 days, we see 58 active pull requests from 38 different contributors, what is very good. :slight_smile:

(Charlotte Morrison) #34


I chose tensorflow for machine learning. This is a very active project and an area that is being developed. When I click on the insights there is a lot of information to unpack. I changed the date to reflect the past month and there were 306 pull requests and 720 active issues. This shows that it is currently active. The most recent pull request was closed 3 hours ago and there are 245 open. There are also over 1500 open issues that need to be worked on.

There is a ton of information that I can use to monitor and assess collaboration in group work. Students like to have the strongest group member do the heavy lifting and claim it was collaborative. I can use this to track the commits the students make and verify they have been working at home because I can see what the student is contributing. As projects progress, the students can use github to create issues and then they can select the ones to complete. This can allow the better coder in the group to be in charge of approving the pull requests to ensure that everyone is contributing good code and they aren’t breaking each others work. This will be very helpful.

Charlotte Morrison

(Dadremarie) #35

There are great tools. I took a couple of shots from insights. Many commits and many merges. So cool. thanks.

(Boydensuperstar) #36

Looking at SmartThings, you can tell by the code frequency and commits. I’d use this to monitor student work on how often they are working on their projects. This will also let me know which students are waiting until the last minute to do work.

(Bill Montana) #37

tensorflow/cleverhans is a current project. There is activity for the last 24 hours. Contributions, commits, and code frequency all show recent activity.




Code frequency

Reflection for classroom use
This is a great tool for assessing collaboration. It answers so many questions with current data. Assessing student work is one of the most time-consuming activities for teachers of programming languages. This can shorten our work and enable us to give better feedback.


(Ayush Bhardwaj) #38

I Found Tensorflow as a live open source project.

**This project has : **

  • 66,000+ Forks
  • 37,000+ Commits (The latest commit was done a few minutes ago.)
  • Around 200 issues and 85 Pull Requests.

  • The contributions graph clearly shows how active this project is.

(Kaushlendra Pratap) #39

Tensorflow is the live project i explored.

**This project has : **

(Amy Dickens) #40

Live Open Source Project :: the repository

The repo is a great example of a living open source project.

The two pieces of evidence that show this are:

  • The last PR was merged just 23 hours ago and the history shows PRs being merged every few days.
  • It has 5,539 stars :star:

Having insights like this is super helpful for assessing the life span of a project - looking over the commit history for example I can see who is contributing to the project most often.
On a public open source project this info will help me to understand who is a maintainer and who is best to help me understand the idiosyncrasies of the project.
On a project I am working collaboratively on - I might be able to see those in my group who aren’t committing and check in on them to see if they need any help.
In assessing the dynamic of group work amongst my students I would be able to corroborate student feedback with the activity within the repo to get a bigger picture of that groups overall work ethic.