Open antea04 opened 1 week ago
Quick links (staging server): Site | Admin | Wizard |
---|
Login: ssh owid@staging-site-happiness-ages
Edited: 2024-06-28 13:20:44 UTC Execution time: 3.76 seconds
Hey @antea04 - nice work!
I've made some suggestions, and I think it would be good to clarify whether they are able to share all the country's data before merging this.
To answer your questions:
I made a chart for life satisfaction for each age group, because I did not find another way to easily compare happiness in an age group between countries. Is there a better alternative?
I think perhaps a good way to present this would be to change this chart to a discrete bar chart, this way the differences between the age-groups are more obvious and we often opt for this when the time-series is fairly flat over time, like it is here. (I would also change the labels so they are a bit shorter e.g. '<30 years', let me know if you aren't sure how to do this.)
The scatterplot you've made is also a really nice way of showing this too. I think just having these two charts is plenty, and I would opt not to create individual maps of the age groups.
Since happiness data per age group is not available for all countries, regional aggregates sometimes include different subsets of countries, which can make this chart look weird (e.g. in "World" view). I added a footnote - is that enough?
Do you mean weird in that for 'World' the value for All ages is lower than each of the individual age-groups? I think perhaps you should set a threshold of around 70% coverage for regions to be calculated. I've suggested a way of doing this in the code review. Hopefully, that will help, if not we should reconsider if calculating regions for the age-group data is helpful. If they are able to share more of the data, this shouldn't be a problem.
Is there an easier/ more elegant way for pivoting the data frame in the grapher step?
I've suggested a way in the code review - you can use age_group as an index, which should mean the pivoting is done automatically. You can also use Jinja templates to give specific metadata to each age-group, when it is included in the index.
The source for the happiness data by age shows up as "Wellbeing Research Center (2024)", instead of "World Happiness Report (2024)" (like other graphs) because they are the producer of the 2024 report, which features this data. Is this discrepency confusing? I made that choice because previous reports have been produced by the "Sustainable Development Solutions Network", so for data which combines multiple World Happiness reports I just gave the source as "World Happiness Report (2012 - 2024)". I felt like this was the simplest solution while respecting the etl defaults where possible, but I can adjust it if its confusing.
I think what you suggest is a good solution!
(see also owid/owid-issues#1581)
Adding age group specific happiness data. Data comes from a .xls file I received from the World Happiness report team, it's the basis of Figure 2.1 here.
Tasks:
Questions for PR Review: