You are on page 1of 9

Big Data Analytics

Sentiment Analysis
Group 1
Problem Statement

● To understand the inherent sentiment behind the Management and Discussion Analysis shared by companies
in each year

● To analyze the effect of external economic factors on the annual reports shared by the companies

Sentiment Analysis

● Classification of data into classes based on tone or emotion ranging from positive to negative

● Branch of NLP that helps companies categorise opinions, reviews, etc. and take action accordingly
Methodology

Data Collection Pre-Processing Analysis Inference

Using Sentiment
analysis tools to
MD&A Report has Elimination of understand the Inferring the
been extracted unwanted rows overall sentiment conclusions based
from All company’s from the dataset; of the Industry and on the acquired
individual annual Segregated the also performing results from the
reports through the dataset into two comparative analysis
years subsets analysis on data
pre and post 2008
Results

Till 2008
Results

Post 2008
Results
Till 2008
Results
Post 2008
Results
Inference
● We observe that polarity is higher in the post 2008 data. This could be due to the uncertainty in the
economic situation post the financial crisis

● In both datasets, trust factor is the highest emotion shown. MDA data is targeted at investors and
stakeholders so its essential to show a trust factor in the annual reports

● There are only minor changes in the other emotions between the two datasets

You might also like