run.py

import numpy as np
import matplotlib.pyplot as plt

# Declaring two empty lists. One for words and one for their corresponding vector representation

words = []
vecs = []

# Getting word and corresponding vector from each line of the model.vec file generated by fasttext 

import codecs # To open the file in specific mode

with codecs.open('model.vec', 'r', 'utf-8') as f_in:
    vocabulary, wv = zip(*[line.strip().split(' ', 1) for line in f_in])
    
 
# Populating the two lists. Need to convert vector values from string to numpy array 

for i in range(10,len(vocabulary)): # Usually skip first 10 words becuase they might be garbage values.
    words.append(vocabulary[i])
    x = wv[i]
    vecs.append(np.fromstring(x,dtype='float32',sep=' '))   
    # np.fromstring takes string values and converts to float32 with space as a separator



# Carrying out Singular Value Decomposition

U, s, Vh = np.linalg.svd(vecs,full_matrices=False) 


# Plotting words and their vector representations

for i in range(len(words)):
        fig = plt.gcf()
        fig.set_size_inches(18.5, 10.5)
        plt.text(U[i,0], U[i,1], words[i])
        plt.xlim((-0.5,0.5))
        plt.ylim((-0.5,0.5))
        
plt.savefig('viz.jpg')