I am very new in this field and recently I have an excercise to build a model. I am so confused whether to build regression model or classification model. What should I do, should I continue on this model or build a new one? And if I continue, what should be improved?
Link to the dataset

I tried to build a random forest model and get both accuracy score and f1-score of 0.5. Features include Region, Country, Item Type, Sales Channel, Order Priority, target is units sold. I tried to include importances of each features to units sold.