SENTIMENT ANALYSIS FOR MARKETING

PREPARED BY: GNANAPRAKASAM .M

Step 1: Data Collection93

You can start by collecting a dataset containing customer reviews of competitor

products. You can consider sources like e-commerce websites, social media platforms, or
review websites. Websites like Amazon, Yelp, or Twitter can be good places to start.
Ensure that the dataset includes text reviews and corresponding sentiment labels
(positive, negative, neutral).

Step 2: Data Preprocessing

Before performing sentiment analysis, it's crucial to clean and preprocess the
textual data:

Remove special characters, HTML tags, and punctuation.

Tokenize the text into words or subwords.
Convert text to lowercase.
Remove stop words (common words like "the," "and," "is" that don't carry much
meaning).
Lemmatize or stem words to their base forms.

Step 3: Sentiment Analysis Techniques

You can employ various NLP techniques for sentiment analysis:

Bag of Words (BoW): Convert the preprocessed text into a matrix of word
counts. Each document (review) is represented as a vector of word frequencies.
 • Word Embeddings (Word2Vec, GloVe): Use pre-trained word embeddings to
represent words in a continuous vector space. Average or concatenate word
embeddings for each document.

Transformer Models (e.g., BERT, GPT-3):

• Utilize pre-trained transformer-based models to capture contextual information in the
text. Fine-tune the model on your sentiment analysis task.

Step 4: Feature Extraction

• Extract features and sentiments from the text data:

• Use the chosen technique (BoW, Word Embeddings, or Transformer) to represent the
reviews as numerical features.
• Apply sentiment analysis algorithms or models to classify each review into positive,
negative, or neutral sentiments.

Step 5: Visualization

• Create visualizations to depict sentiment distribution and analyze trends:

• Generate bar charts or pie charts to show the distribution of positive, negative, and
neutral sentiments.
• Create word clouds to visualize the most frequently mentioned words in positive and
negative reviews.
• Plot time series graphs to observe how sentiments change over time.

Step 6: Insights Generation

• Extract meaningful insights from the sentiment analysis results to guide business
decisions and elaborate on the content:

Identify common themes or topics in positive and negative reviews.

• Analyse the sentiment trends over time to see if there are any seasonal patterns or
product-specific events affecting sentiment.
Compare sentiment scores across different competitor products.

• Look for keywords or phrases that frequently appear in positive or negative

reviews, which can provide actionable insights for product improvement.

• By following these steps, you can perform sentiment analysis on customer reviews
of competitor products, gain valuable insights, and make informed business
decisions based on customer feedback.
PROGRAM:
from mpl_toolkits.mplot3d import Axes3D
from sklearn.preprocessing import StandardScaler
import matplotlib.pyplot as plt # plotting
import numpy as np # linear algebra
import os # accessing directory structure
import pandas as pd # data processing, CSV file I/O (e.g. pd.read
_csv)
There is 1 csv file in the current version of the dataset:
In [2]:
for dirname, _, filenames in os.walk('/kaggle/input'):
for filename in filenames:
print(os.path.join(dirname, filename))
/kaggle/input/car data.csv
/kaggle/input/cardata.R
The next hidden code cells define functions for plotting data. Click on the "Code" button
in the published kernel to reveal the hidden code.
unfold_lessHide code
In [3]:
# Distribution graphs (histogram/bar graph) of column data
def plotPerColumnDistribution(df, nGraphShown, nGraphPerRow):
nunique = df.nunique()
df = df[[col for col in df if nunique[col] > 1 and nunique[c
ol] < 50]] # For displaying purposes, pick columns that have betw
een 1 and 50 unique values
nRow, nCol = df.shape
 columnNames = list(df)
nGraphRow = (nCol + nGraphPerRow - 1) / nGraphPerRow
plt.figure(num = None, figsize = (6 * nGraphPerRow, 8 * nGra
phRow), dpi = 80, facecolor = 'w', edgecolor = 'k')
for i in range(min(nCol, nGraphShown)):
plt.subplot(nGraphRow, nGraphPerRow, i + 1)
columnDf = df.iloc[:, i]
if (not np.issubdtype(type(columnDf.iloc[0]), np.number)
):
valueCounts = columnDf.value_counts()
valueCounts.plot.bar()
else:
columnDf.hist()
plt.ylabel('counts')
plt.xticks(rotation = 90)
plt.title(f'{columnNames[i]} (column {i})')
plt.tight_layout(pad = 1.0, w_pad = 1.0, h_pad = 1.0)
plt.show()
unfold_lessHide code
In [4]:
# Correlation matrix
def plotCorrelationMatrix(df, graphWidth):
filename = df.dataframeName
df = df.dropna('columns') # drop columns with NaN
df = df[[col for col in df if df[col].nunique() > 1]] # keep
columns where there are more than 1 unique values
if df.shape[1] < 2:
print(f'No correlation plots shown: The number of non-Na
N or constant columns ({df.shape[1]}) is less than 2')
return
corr = df.corr()
plt.figure(num=None, figsize=(graphWidth, graphWidth), dpi=8
0, facecolor='w', edgecolor='k')
corrMat = plt.matshow(corr, fignum = 1)
plt.xticks(range(len(corr.columns)), corr.columns, rotation=
90)
plt.yticks(range(len(corr.columns)), corr.columns)
plt.gca().xaxis.tick_bottom()
 plt.colorbar(corrMat)
plt.title(f'Correlation Matrix for {filename}', fontsize=15)
plt.show()
unfold_lessHide code
In [5]:
# Scatter and density plots
def plotScatterMatrix(df, plotSize, textSize):
df = df.select_dtypes(include =[np.number]) # keep only nume
rical columns
# Remove rows and columns that would lead to df being singula
r
df = df.dropna('columns')
df = df[[col for col in df if df[col].nunique() > 1]] # keep
columns where there are more than 1 unique values
columnNames = list(df)
if len(columnNames) > 10: # reduce the number of columns for
matrix inversion of kernel density plots
columnNames = columnNames[:10]
df = df[columnNames]
ax = pd.plotting.scatter_matrix(df, alpha=0.75, figsize=[plotSize, plotSize], diagonal='
kde')
corrs = df.corr().values
for i, j in zip(*plt.np.triu_indices_from(ax, k = 1)):
ax[i, j].annotate('Corr. coef = %.3f' % corrs[i, j], (0.
8, 0.2), xycoords='axes fraction', ha='center', va='center', siz
e=textSize)
plt.suptitle('Scatter and Density Plot')
plt.show()
Now you're ready to read in the data and use the plotting functions to visualize the
data.
Let's check 1st file: /kaggle/input/car data.csv
In [6]:
nRowsRead = 1000 # specify 'None' if want to read whole file
# car data.csv may have more rows in reality, but we are only loa
ding/previewing the first 1000 rows
df1 = pd.read_csv('/kaggle/input/car data.csv', delimiter=',', n
rows = nRowsRead)
df1.dataframeName = 'car data.csv'
nRow, nCol = df1.shape
print(f'There are {nRow} rows and {nCol} columns')
There are 301 rows and 10 columns
Let's take a quick look at what the data looks like:
In [7]:
df1.head(5)

Out[7]:

Car_Name company Year Selling_Price Present_Price Kms_Driven Fuel_Type Seller_Type Transmission Owner

maruti
0 Ritz 2014 3.35 5.59 27000 Petrol Dealer Manual 0
suzuki

maruti
1 sx4 2013 4.75 9.54 43000 Diesel Dealer Manual 0
suzuki

maruti
2 Ciaz 2017 7.25 9.85 6900 Petrol Dealer Manual 0
suzuki

maruti
3 wagon r 2011 2.85 4.15 5200 Petrol Dealer Manual 0
suzuki

maruti
4 Swift 2014 4.60 6.87 42450 Diesel Dealer Manual 0
suzuki

Distribution graphs (histogram/bar graph) of sampled columns:

In [8]:
plotPerColumnDistribution(df1, 10, 5)

Correlation matrix:
plotCorrelationMatrix(df1, 8) In [9]:
Scatter and density plots:

In [10]:
plotScatterMatrix(df1, 15, 10)
Conclusion:
✓ This concludes your starter analysis! To go forward from here,
click the blue "Fork Notebook" button at the top of this kernel.

✓ This will create a copy of the code and environment for you to
edit. Delete, modify, and add code as you please.