I found the data today pretty tricky to work with for a few reasons - chiefly, that each country reports its food waste differently, making meaningful comparison between countries difficult. Additionally, different years yielded hugely different reported data even within the same country and so comparison over time was not great either. Finally, it was not obvious when totals or sub-totals were included in the data itself and so aggregating values based on the existing fields would end up in a lot of duplication.
With that in mind I tried to pull out the most useful data that I could. After some digging, I noticed that in the 'Variable' field both the US and Canada split their food waste data up into similar food types and so decided to run with those two countries.
For my flow I filtered onto only the United States and Canada and used formulas to convert all of the values to tonnes (some were kilograms and some were million kilograms). Using the text to columns tool, I then split out the 'Variable' field using a dash to separate out the food type and then filtered out any rows that were not classed as 'food loss'.
The output of this flow gave me data for Canada and the US detailing the food waste across a variety different food types spanning several years. I used groups in tableau to group similar food types together that were reported differently across the years and countries.
Once I had this data into tableau and grouped correctly it was relatively easy for me to construct the charts that ended up in the final dashboard. I took population data to calculate the tonnes of food wasted per person and chose to compare 1995 to 2010 due to those years having the most similar data.
Using parameter actions the user can change which country's data is displayed in the big numbers and charts on the bottom half of the viz.
The final dashboard can be found here.