Scribd
Upload
61 views6 pages

Sentiment Analysis Meeting Insights

The document provides a sentiment analysis report of a meeting with visualizations of sentiment distribution, individual speaker analysis, and meeting dynamics. It compares rule-based and neural network approaches to sentiment analysis and describes the Python code and libraries used to generate the sentiment analysis and visualizations.

Uploaded by

pythondev.erginous
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Topics covered

  • Analysis Report,
  • Pie Charts,
  • OpenAI API,
  • Speaker Analysis,
  • Seaborn,
  • Negative Sentiments,
  • Deep Learning,
  • Data Visualization,
  • Python Programming,
  • Visualization Tools
61 views6 pages

Sentiment Analysis Meeting Insights

The document provides a sentiment analysis report of a meeting with visualizations of sentiment distribution, individual speaker analysis, and meeting dynamics. It compares rule-based and neural network approaches to sentiment analysis and describes the Python code and libraries used to generate the sentiment analysis and visualizations.

Uploaded by

pythondev.erginous
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Topics covered

  • Analysis Report,
  • Pie Charts,
  • OpenAI API,
  • Speaker Analysis,
  • Seaborn,
  • Negative Sentiments,
  • Deep Learning,
  • Data Visualization,
  • Python Programming,
  • Visualization Tools

Sentiment Analysis Report

About this Report


Conducting sentiment analysis on different speakers in a meeting signifies a nuanced
understanding of the emotional dynamics within the discussion. By evaluating
sentiments expressed by individual speakers, we analyze patterns in communication
styles, collaboration, and potential areas of concern. The resulting visualizations,
such as pie charts and count plots, provide a comprehensive overview of the
meeting's emotional tone. Essentially, analysis transforms spoken words into valuable
insights, aiding in effective collaboration, team management, and overall meeting
effectiveness.

Here are some key points that your report may signify:

Sentiment Distribution:
The pie chart can provide an overall view of the sentiment distribution in the meeting.
We have included sentiments such as Neutral,Slightly Positive,Slightly Negative ,
Moderate Positive , Moderate Negative , Strong Positive , Strong Negative .

Individual Speaker Analysis:


By creating count plots based on speakers and sentiments, we identify which
participants expressed more positive or negative sentiments. This could be useful for
understanding the overall mood or contributions of specific individuals.

Meeting Dynamics:
Patterns in sentiment changes throughout the meeting may reveal the dynamics of the
discussion. For example, we might notice shifts in sentiment during specific topics or
when certain individuals speak.

Identifying Concerns or Issues:


Negative sentiments may highlight areas of concern or disagreement. This could be
important for addressing issues that may need resolution or further discussion.

Positive Aspects:
On the flip side, positive sentiments can highlight successful collaboration,
agreement on key points, or a generally positive atmosphere within the meeting.
Comparison between VADI generated Sentiments and OpenAI generated
Sentiments.

VADI generated sentiments are based on Rule based system using libraries such as
NLTK etc. Whereas OpenAI based sentiment analysis considers the flow of the
conversation . It is obvious that OpenAI model is trained on a large amout of data, it
has a capability to mimic behaviour like a human being. So , our approach is based
on Neural Networks.(Deep learning based approach).

The sentiments are stored in a comma separated file (.csv) with speaker, statement
and sentiment. After that visualizations are plotted such as pie chart and countplot for
better understanding .

Below are the visualizations :


All these visualization are based on the conversation done by two people and we have
manually verified everything.

Technology Used:
Programming language: Python
API: OpenAI API
Framework: Langchain
Libraries :
Pandas (For creating DataFrame)
Matplotlib (For visualization)
Seaborn(For advanced Visualization)

Code just takes text file containing speaker:transcription. Using LangChain + OpenAI
+ Prompt Engineering , It firstly generates the sentiments based on each statement by
speaker and then save it in DataFrame along with Speaker , statements. DataFrame
then saved locally as .csv file. After that all visualizations are plotted using libraries
for better understanding.

You might also like