Dashboard week: US consumption of vegetables

by Jonatan Raphael-Amanrich

The challenge of the day for our dashboard week consisted in creating a viz about the US consumption of vegetables. We first needed to create a sketch of our dashboard using only one sheet of the dataset that we were going to use, then we needed to import the data on Tableau prep and do the necessary cleaning and preparation and finally we had to crate our viz.

The main issue faced during the preparation was the large amount of sheets needed to be imported and the fact that these sheets had 2 headers for our values which created some issues in tableau prep. After spending some time looking for a solution, we realised that using the ‘cleaned with data interpreter’ option in prep was actually proceeding to a pre-cleaning and removing the extra headers which made the end of the preparation way easier. Indeed, I just had to remove all the fields that would not be used for the dashboard, filter out some vegetables as they referred to broad categories like legumes instead of specific vegetables or things like ‘potato chips’ which can hardly be considered vegetables, at least outside of the U.S. I then used a 2 calculated fields, one to get the type of conservation of each vegetable (fresh, dry, frozen…) and I used a regex expression to get a field containing each individual vegetable. Everything was ready to get exported to Tableau.

See my workflow below:

For my dashboard, the idea was to create a graph showing the evolution of the consumption of vegetables over time, adding a filter to show specific vegetables. I also wanted to create a Top N vegetable consumed for a specific year and create a drill down showing the consumption values per conservation type. I also created a pie chart showing the percentage represented by every type of conservation for the vegetable and year selected. Apart from time limitation which was the major issue, I faced many challenges that I was unable to overcome. First, there was conflicts between my filters and parameters and it took me quite long to figure out what was going but I succeeded in sorting them out in the end. Then, I tried to create a drill-down bar showing the consumption per conservation type when selecting a vegetable but was unable to make it work.

Final thoughts about this challenge: It was really hard to stay focus on the task and keep everything organised with such a short deadline so I had a few times during the day where I was feeling pretty lost and not remembering exactly what was my end goal after starting a task. Despite all these challenges I have created a dashboard that seems to be functional even though a bit limited in terms of insights I have to admit.


Jonatan Raphael-Amanrich