Personal project in R: analyze Kaggle dataset on Coursera features and visualize findings with Tableau

Hi everyone!

If it is ok (it does not seem to go against any of the forum’s rules, but please delete if it does!) I would like to share with you the project I made as final project for a Coursera introductory course on data analytics. The code is in R langague, the commentary is a markdown notebook (.rmd) and the visualizations are made with Tableau public.

I collected two datasets from Kaggle: one describes the characteristics (content, duration, difficulty level, language, and so on) of hundreds of online courses on the Coursera platform; the other described the results of the Coursera Global Skills report 2021
capstone.zip (153.5 KB)
, that ranks different countries and world region based on the users’ ability to complete the courses.
I analyzed data from the two sets in order to understand whether there is any relationship between the courses’ features and the users’ abilities, and how it varies across different world regions.

This is my first project ever, so I am full of doubts about everything:

  • Is my code good? Is it too verbose, or involuted, or simply wrong?

  • Does my analysis make sense? All the choices I made as to what data to discard, what to keep, how to transform them, how to use them to answer my questions…are they sensible choices?

  • Are my visualizations clear? It was also my first work with Tableau so I feel like much of it is kind of ‘forced into place’.

  • Is my notebook comment too long, too detailed, or else too obscure?

I know it’s long so I do not expect anyone to actually read it, but if someone could just skim through it and give me any kind of feedback, it would mean the world to me, even because I’d like to improve it and insert it in my portfolio for future potential employers.

Thank you so much. Hereby I attach:
the link to my code on Kaggle (but without the whole commentary) and the link to the original datasets:

The link to my Tableau visualizations: https://public.tableau.com/app/profile/federica.minozzi/viz/capstone_16620195324340/Dashboard1

And the zip file containing the .rmd file with the whole commentary for the code, and the data files.

1 Like

Hi @fedeminozzi ,

Great job finishing your tableau dashboard. I was not able to read the Kaggle notebook but I did have the opportunity to review your tableau dashboard. Great job adding the interactive controls for views and skill area charts. Great interactive chart titles that change when a new view is selected. I enjoyed learning the different facts about Coursera. I was surprised to learn the majority of classes are beginner level classes with very few advanced level classes. Great job creating meaningful visualizations.

My only suggestions deal with adding more explanations and labels so the charts are easier to understand:

1. * Under the title "Coursera courses features and global proficiency: is there a relation?":

  • you may want to add a subtitle to identify the exact course features you identified
  • you may want to add a short explanation of how you are measuring global proficiency
  1. Did you think about a cluster chart to show correlation since you posed the question if there was a relation

  2. Are the drop down options in the select view box features… if so you may want to change name to “Select a Feature”

  3. It was a little confusing in the beginning to follow which controller changed which chart so if you can move the controller closer to the chart… it would make it easier and quicker to follow

  4. You may want to add short explanation about the x-axis scale for “Regions” and “Mapped Skill” chart

  5. You may want to add a short explanation about the proficiency line and proficiency scale

  6. You may want to add a short explanation of how each chart answers your first question - is there is a relation between course features and global proficiency

These are just my personal thoughts. I enjoyed learning more about Coursera platform. Great job finishing your first project!

2 Likes

HI, thanks for the helpful suggestions!

I agree that the dashboard should be more descriptive and explanatory. The problem is, I can’t fit anything else into the allotted space. I do not want to reduce the plots’ sizes because they are already so dense and I do not want them to become unreadable, but if I try to add anything else it becomes a mess.
Maybe dashboard is not a good option to display so many plots, or maybe I should select only one or two plots and discard the others (or make two separate dashboards), or perhaps there are settings that could help me but I don’t know how to use them (any chance you got suggestions for me in this sense?)

I had not thought of a cluster chart, it is an interesting idea to explore my data, and also I agree that it would be helpful to place the controllers in a way that is more clearly related to each chart. I’ll make these changes straight away.

Thank you so much for your precious feedback!

Hi @fedeminozzi ,

You may want to add hover text help where you can add information that will only display when you hover over a topic. The below screenshot was taken from tableau public visualization of the day.
screenshot

Put Some Prep In Your Step Channel Analytics | VOTD

by[Jared Flores]

Click here to read whitepaper on “10 best practices for building effective dashboards”.
Click here to read additional tips on creating dashboards.
Here is a link to a tableau public blog post discussing hover help menu. You can also click here to view additional public tableau dashboards.

2 Likes

Hi, thanks for the helpful suggestions and useful links. Hover text could definitely solve my problem without me needing to remove some of the plots, and it also looks cool! Thank you so much!

1 Like