試試

df.pivot('name', 'subject', 'score').plot.bar()

基本pandas 視覺化原則是，使用pivot 或melt搞出製圖所需的表格格式。

下列是我跑的程式碼(Jupyter Notebook)

% matplotlib inline
data = [ {"name":"abc", "subject":"A", "score":40}, {"name":"abc", "subject":"B", "score":60}, {"name":"abc", "subject":"C", "score":40},
        {"name":"xyz", "subject":"A", "score":10}, {"name":"xyz", "subject":"B", "score":90}, {"name":"xyz", "subject":"C", "score":30}]
df = pd.DataFrame(data)
df.pivot('name', 'subject', 'score').plot.bar()

結果：