Skip to content

Latest commit

 

History

History
36 lines (30 loc) · 2.35 KB

README.md

File metadata and controls

36 lines (30 loc) · 2.35 KB
Raw

Sentiment-Analysis

This is a project on Sentiment Analysis based on Python 3.8. It was made by Deepesha Burse as a personal project. With this project, I have analysed the reviews of a women's clothing company. The data has been extracted from a CSV file.

Packages Used:

  1. pandas
  2. numpy
  3. seaborn
  4. nltk
  5. wordcloud among others.

Motivation

In order to enhance myself, I had done two courses, Launching into Machine Learning and Introduction to Data Science in Python. I enjoyed learning these concepts and hence wanted to try my hand at a project based on them. I came across a project on sentiment analysis, and felt like this would be the perfect project to reinforce everything I have learnt.

Working of Project:

1. Extracting Data

I extracted the data from a CSV file that contained the reviews along with some other order details using pandas. This data got stored in the form of a DataFrame.

2. Data Visualization

In this step, I used the package seaborn to plot the ratings and polarity ratings on graphs. It also included data preprocessing.

3. Data Cleaning

This stage included:

  • Dropping unnecessary columns
  • Dropping rows containing missing values (NaN values)
  • Tokenization — convert sentences to words
  • Removing unnecessary punctuation, tags
  • Lowering text
  • Removing stop words — frequent words such as "the", "is", etc. that do not have specific semantic
  • Stemming — words are reduced to a root by removing inflection through dropping unnecessary characters, usually a suffix.
  • Lemmatization — Another approach to remove inflection by determining the part of speech and utilizing detailed database of the language.
  • Bag of Words
  • TF-IDF

4. Sentiment Analysis

Taking this cleaned data, we will perform Sentiment Analysis using the Sentiment Intensity Analyzer. Then, we can extract the pos, neg, neu, compound values. Following this, I applied the doc2vec and TFIDF Vectorizer and added features such as number of words and number of characters. I, later made a wordcloud of the top words found in the reviews and listed down the top positive and negative reviews.