For the individual visualization project, the minimum number of graphs is one.
Of course it should be a really excellent graph, and it might be a figure with multiple subplots in it!
Ideally, your one graph would be able provide a lot of good information about your dataset, and show your audience multiple aspects of your at the same time (without overwhelming them).
It's also fine to have several graphs, but they should be thematically related to each other in telling a coherent story about your data. (Definitely DO NOT make 20 different unrelated graphs from various columns of your data, and expect your readers to make their own sense of it!)
Here's a website discussing some guidelines for good data visualization:
http://paldhous.github.io/ucb/2016/dataviz/week2.html
