Open foscanit opened 9 months ago
🐬📊Congrats on your project!
README Your README is wel organized and very informative. I really like how you detailed your process, the questions you intended to answer and a very good explanation about your data sources. One way you could improve this README, you could add images of your visualizations right alongside your analysis, to give the full picture right here in your repository for those readers that don’t click the link to Tableau.
Code The comments and organization of your notebooks is very good. Keep working like that. And the error handling on your functions is always welcome. To take them to the next level, you can add docstrings with the goal of the function and explanations of inputs and output.
SQL Your queries are well done and you show a great understanding of how to get the results you need to answer your questions. It is also very good that you include the hypothesis you are trying to answer with each query plus comments about what you are doing.
Tableau The way you organized your story, with one story point for each question is great. It makes navigation the information easier.
In cases like the scatterplot of population vs registered languages, where you have two very clear outliers (China and India) it may be a good idea to make a graph without them so your x axis goes from 0 to 350 M and the data points are not all concentrated to the left of the graph. If you decide to go that way, you always have to add a clarification of the countries that were left out and include their particular values for each variable. Or even make the complete graph available as an extra.
When you have more than 4 or 5 categories, like in your Languages and dialects spoken tree map, it is easier to interpret if you show the name of the country directly in each rectangle, instead of your reader trying to match the color of the legend to each square.
And it is best to be consistent: if you don’t have percentages for all countries, it is better to leave all percentages out of your graph. You can mention them in your analysis. In your visualization, only Palau and Marshall Islands has that description by percentage.
The way you did your analysis with top 10 and bottom 10 countries to compare is very engaging for a reader: everybody likes a ranking. I just have a question: did you create scatter plots with all the countries and the variables you wanted to analyze? Maybe the relationships you found in the top and bottom are also present in the middle and you can generalize your analysis or maybe there are other types of relationships you could detect in the middle (maybe there isn’t, but the best idea is to check).
Sorry, I had my project almost finished since last Tuesday, with Tableau's visualizations, but the Readme was still pending....
https://github.com/foscanit/project-4-sql-tableau