Why not just to make a stacked bar plot? Somethin like this:

res = df.T.drop_duplicates().T
# df here is your source dataframe
fig, ax = plt.subplots()

for name, group in res.groupby("PVR Group"):    
    ax.bar((group['min_x']+(group['max_x']-group['min_x'])/2),
           group['max_y']-group['min_y'],
           width=(group['max_x']-group['min_x']),
           bottom=group['min_y'],
           label=name,
           edgecolor="black")

ax.legend()
plt.show()

I believe you could adjust the design to meet your taste.