hannesdatta / course-dprep Goto Github PK

e.g., simplify code

• generating markdown docs, also from the command line; PDF & HTML
• copy-pasting code snippets from the web and use them
• replicate for different country

• Think carefully about self paced tutorials versus Q and a sessions. There should not be extra material in these tutorials, but a bit of more structured preparation may be necessary.
• Include official GitHub On boarding exercises.

Especially with regard to assignments

filter required?
aggregation required? (e.g., new variables that can be computed)
recoding required?
merging required?
operationalization required?/new variable required?
etc.

plotting:

• axis labeling
• title
• notes
• ...

make it a check and discuss (in the lecture) how to go from raw to prepared.

• Make more knowledge clips of stuff relevant for a tutorial
• Make tutorials "mandatory" with a deadline or social proof
• "Pulse" - social proof app
  • indicate how far you are
  • possibility to ask questions/log where you are stuck
  • participation requirement (bonus point system)
• Replace live streams with Q&As
  • go through questions that were shared before

• verhuizen tisem airbnb v.2 naar TSH (organization: tilburgsciencehub)
• git lfs implementeren
• nodige bestanden commiten met Git LFS (1GB)

• Online Python frameworks: https://repl.it/site/ide, https://www.pythonanywhere.com, https://colab.research.google.com/notebooks/intro.ipynb#recent=true

more in ppt in course folder on surf

• research motivation/context rather than RQ
• use issues (+ point to relevant page on TSH)
• use the project board (+ point to example)
• restructure scrum article
• solve issues: git pull/push + most important git problems
• dissemination opportunities: not so much done yet + explain how this course is different

Incorporate more exercises - Student have requested this and the feedback. One way to do this is to have short questions, e.g. to scrape some thing at the end of the lecture. Or to solve certain problems.

• "proof of investment in skills" (submitted tutorials/data challenges on Canvas)
• written feedback to team members
• assessment of own performance and that of team members

Ideas:

• months before: check osiris
• before the course starts (check enrollments, post announcement), etc.

Estimate number of hours for entire course, Using the template I have received from the program coordinator. Accordingly, reduce the workload.

Overview needs to be by

• module (including preparation before the course starts, up to including the exam prep & exam)
• item-by-item (see course modules)

Inspired by this template, please adjust the syllabus to the format (style headings, etc.) of odcm's syllabus.

https://dprep.hannesdatta.com/docs/course/syllabus/

...or an open source alternative, little more data friendly...

• https://books.google.nl/books/about/Data_Wrangling_with_Python.html?id=qgyLDwAAQBAJ&printsec=frontcover&source=kp_read_button&redir_esc=y#v=onepage&q&f=false
  https://cengel.github.io/R-data-wrangling/index.html
• https://medium.com/@ODSC/top-data-wrangling-skills-required-for-data-scientists-8a6b7dc604a7
• https://www.edx.org/professional-certificate/harvardx-data-science
• https://www.topbots.com/data-preparation-for-machine-learning/
• https://www.trifacta.com/data-preparation/
• https://www.oreilly.com/openbook/make3/book/index.csp

Evaluate whether this may make sense adding to TSH: https://books.ropensci.org/targets/

Deliverable:

• Tell us whether this should be a new building block at TSH, or merely a link somewhere on our automation sites on TSH.
• If not a building block, add the link at relevant places on Tilburg Science Hub, and the dprep course.

Teaching activities

• Wellicht benoemen hoe dit vak zich weerhoudt tot RSM en oDCM (notes)

About you / Covid-19

• Vaak is de response rate een stuk hoger bij het gebruik van een poll software. Daarom wellicht een idee om het om te zetten in een aantal korte vragen (bijv. met sli.do):
  • What's your undergraduate degree? (business administration, marketing, economics, engineering, other)
  • What stage are you in? (beginner, intermediate, advanced)
  • What are your long-term career ambitions (go into industry, do a PhD)?
  • Where are you at right now? (at home, at a friend's place, abroad, etc.)
  • Home office, on a scale from 1 to 10? (1, 2, ..., 10)
  • Wanna have the last half-an-hour of this with a borrel? ;) (yes, yes)

Can't find stuff

• Discuss other common issues the course addresses
  • Lost track of files
  • Excel becomes slow (large datasets)
• Add concrete examples (screenshot of messy directory of past research projects, report.pdf -> final_report.pdf -> FINAL_final_report.pdf, etc. )

What's a computation-intensive project

• Perhaps make a little bit more concrete (for students who never programmed in R)
• For example, running this analysis for my paper would take X days on my local machine (and forces me to keep my computer running 24/7). Thanks to AWS I do the same in X hours.

Course framework

• Maybe it makes more sense if you first show the course framework and then the team project?

Canvas versus the web

• Sign up for teams through Canvas?

Comments on coding

• Mention how we aim to remedy this initial hurdle? (quick feedback loops especially in the first few weeks)

Help

• Give concrete example on how they can find help with Google and Stackoverflow (say that I don't know how to do X in R) -> prepare them for what's coming up.
• Go over the help workflow:
  1. Read documentation
  2. Google / Stackoverflow
  3. Ask a friend
  4. Google / Stackoverflow
  5. Ask a friend
  6. Ask course instructor (when to use which channel, guidelines on what to include; the better the question the better the answer)

Missing

• Briefly touch upon the grading scheme for the course (couldn't fit it in this slide deck)
• Live coding sessions where they can help each other out (what to expect, etc.)
• Set the scene for what's coming next:
  • Show them where they can find what to do each week (click through it), discuss learn -> practice -> create
  • How to right click, download file as...
  • Briefly mention the difference between R and RMarkdown files
  • Manage expectations: DataCamp courses are simplistic (copy-paste etc.) and should not take too long, the data challenges are more challenging and will likely take more time, and are more representative of the level we expect in this course.

interactive R sheets for practice questions

Hi @andreantonacci, is there a quick way to bring back search on my open edu classes? Would be cool if you could try it out in a feature branch. Thanks!

Ideas for Git:

will supply w/ excel sheet w/ new planning