sum of orders would be "the total number of pizzas ordered", this is sum of total, which is price x orders = revenue. so the chart shows not the popularity but how much money each one brings

again, this is revenue, not orders let's turn 01234 into human readable days? you probably know how, but if not, a few others did solve it, ask around

price per unit depends a lot on size. i suggest splitting name column to name and size columns and analyse sizes as facet Y or names as facet Y independently

i think the dips by the end of year are because some sizes were temporarily discontinued