semio / ddf--gapminder--systema_globalis Goto Github PK

The SG dataset is created by merging new data into the old GW dataset. But some of the indicators may not be appropriate to deep merge with the old data(keep the union of datapoints in both old and new), because the new data may have different collecting/calculating method.

For example an indicator calculated with 2011 usd is not good to deep merge with updated version calculated with 2016 usd. We should overwrite the 2011 version with 2016 version in this case.

need to find out which ones should use deep merge and update recipes

dp removed: brn 2008 5.5

Because of deep merge, there should not be removal...

cumulative_co2_emissions_tonnes:

1. most dps are 1000 times smaller than old one
2. new dps added, but some of them are negative number (aus 1850s)
3. Some countries have 0 added in front of time series (e.g. irn)
4. Because of deep merge some changes in ""about"" here still work: https://docs.google.com/spreadsheets/u/1/d/1sFMSK6RsphnEFEvD6ucUpadA6ghIhfXbW04W754acjM/pub#"

yearly_co2_emissions_1000_tonnes and yearly_co2_emissions_tonnes:

I checked the spreadsheets in graph_settings for both indicators. They both use 1000 metric tonnes for unit. Difference is that this one is 2012 ver. and the other one is 2015 ver.

But in generated SG I set yearly_co2_emissions_1000_tonnes = yearly_co2_emissions_tonnes / 1000. Need to fix this