This project is Sentiment Analysis based on Twitter comments. The code is written in Python and we have used Jupyter Notebook.
Here I have used two Python Libary for Sentiment Analysis and Data Visualization
- Scikit Learn (sklearn)
- Tensorflow
- scattertext & spacy
NOTE: We are using lbfgs solver for Logictic Regrerssion Model. lbfgs stand for: "LimitedMemory-Broyden–Fletcher–Goldfarb–Shanno Algorithm". It is one of the solvers' algorithms provided by Scikit-Learn Library. The term limited-memory simply means it stores only a few vectors that represent the gradients approximation implicitly. It has better convergence on relatively small datasets. So I have used max_iter value as 1000.
We have imported below moduled from sklearn libary.
from sklearn.model_selection import train_test_split # function for splitting data to train and test sets
#feature extraction
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.linear_model import LogisticRegression #Logistic Regression Model
from sklearn.pipeline import Pipeline
from sklearn.metrics import accuracy_score, precision_score, recall_scoreHere we have used LSTM Model (Long Short Term Memory)
We have imported below moduled from tensorflow libary
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Dense, Dropout, SpatialDropout1D, EmbeddingWe have used Scattertext and spacy to create a interactive html
import spacy
import scattertext as st
#creating corpus object
nlp = spacy.load('en_core_web_sm')
corpus = st.CorpusFromPandas(data,
category_col='sentiment',
text_col='text',
nlp=nlp).build()
#creating html file
html = st.produce_scattertext_explorer(
corpus,
category = "Positive",
category_name = "Positive Tweets",
not_category_name = "Negative Tweets",
width_in_pixels = 1000,
metadata = data["candidate"]
)
open('Tweet_Analysis.html','wb').write(html.encode('utf-8'))The final output might look something like this:
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent