In [40]:

    

from pml.api import *
import matplotlib.pyplot as plt


    

In [41]:

    

data = load("../dataset_ext2.csv")
data = data.drop_empty_samples()
data.fill_missing_with_feature_means()


    

In [42]:

    

data.get_label_value_counts()


    

    
Out[42]:





f    30
s    26
p    16

In [43]:

    

# Get first principal component
princomp = pca(data, 1)
princomp.get_first_component_impacts()


    

    
Out[43]:





MATH110    0.472499
CSC111     0.339158
CSC115     0.335191
MECH141    0.320300
PHYS122    0.288521
PHYS125    0.287089
MATH100    0.271468
ELEC199    0.268938
MATH101    0.260933
CHEM150    0.242886
ENGL135    0.089390
ENGR120    0.002989
ENGR110    0.002226

In [44]:

    

princomp.feature_list()


    

    
Out[44]:





[0]

In [45]:

    

# plot configuration
markers = {"s": "o", "p": "^", "f": "x"}
colours = {"s": "g", "p": "y", "f": "r"}


    

In [52]:

    

# Plot 1st PC Weights For Each Sample
plt.figure(1)
plt.subplot(3, 5, 1)

for label in data.get_label_set():
    filtered = data.label_filter(label)
    
    xs = filtered.get_sample_ids()
    ys = princomp.get_rows(xs).get_column(0)

    plt.scatter(xs, ys, color=colours[label], 
                marker=markers[label])

plt.xlabel("Sample Id")
plt.ylabel("First Principal Component")
plt.title("1st PC Weights For Each Sample")


    

    
Out[52]:





<matplotlib.text.Text at 0x9715510>

In [53]:

    

# For each course, plot grades for each sample (student)
for i, course in enumerate(data.feature_list()):
    # Plot 1 used for 1st PC, so start at 2 here
    plt.subplot(3, 5, i + 2)

    for label in data.get_label_set():
        filtered = data.label_filter(label)
    
        xs = filtered.get_sample_ids()
        ys = filtered.get_rows(xs).get_column(course)

        plt.scatter(xs, ys, color=colours[label], 
                    marker=markers[label])

    plt.xlabel("Sample Id")
    plt.ylabel("%s Grade" % course)
    plt.title("%s Grade For Each Sample" % course)
    
plt.show()


    

In [ ]: