SentAnalysis-py : Code & Screenshots


Last week was crazy ! I have been coding for a kernel module for the simulation project and results went a bit unexpected. I rather tried to explore sentiment analysis for a friend that I was helping in his presentations. Its not that complicated after-all for basic thoughts as a skill-up exercise.

You can use the code after :

  1. You have a output.json {contains tweets from streaming API}. Download a sample version here.
  2. Should know that the syntax is in Python 2.7 so wont work with 3.X. Also I am using AFINN and soon would be using Wordnet {In a complicated way}.
  3. Run it after you have all the required import libraries including Json and oauth2.
  4. Run it as $python . I have added few screenshots.


import sys
import json
import re

def hw(sent_file,tweet_file):
sent_dict = {}

for line in sent_file:
line_list = line.split()
if len(line_list) > 2:
length = len(line_list)
temp_line_list = []
temp_line_list.append(" ".join(line_list[:length-1]))
line_list = temp_line_list
sent_dict[line_list[0]] = float(line_list[1])

for line in tweet_file:
## print "a new tweet"
dict = json.loads(line)
sum = 0;
if 'text' in dict.keys():
text = dict['text']
## print text.encode('utf-8')
words = text.split()

for word in words:
word = re.sub('[^0-9a-zA-Z]+', '', word)
sum += sent_dict.get(word, 0)
print sum
def lines(fp):
print str(len(fp.readlines()))

def main():
sent_file = open(sys.argv[1])
tweet_file = open(sys.argv[2])

if __name__ == '__main__':

Sentiments of the Tweet

Lastly, the plot can be done using matplotlib but I used Google charts for fast depiction.








Share Button