Sentiment Analysis of Social Media data for Brands
Business Requirement:
The client is a digital marketing analytics company that helps its client figure out how their brands are doing on social media. It also analysis the various campaigns run by the end clients to study the social media impact and trends.
The client’s requirement was to automate their manual sentiment analysis process.
Nuage Solution:
Nuage developed a Cloud hosted Sentiment analysis platform to analyze social feeds received from different social channels (Facebook, Twitter, LinkedIn, YouTube) primarily comments and tweets.
The main challenge was to handle diverse file formats of incoming feeds. To tackle this, a data ingestion layer was developed to convert external formats into a format that our application could comprehend. An admin UI was created to map incoming formats to the internal data structure, enabling flexibility to handle multiple formats and making the system compatible with various social platforms and future-proof.
Our team works on data cleansing and noise removal using NLTK and other custom pre-processing algorithms such as punctuation removal and stemming to prepare the raw text for mining. This pre-processing phase is essential as it prepares raw text for mining by facilitating information extraction and enabling the application of machine learning algorithms.
To perform sentiment analysis, we transformed the pre-processed data into a machine-readable dataset using the Term Frequency-Inverse Document Frequency (TF-IDF) method.
We used a lexicon-based approach to automate the client’s current processing methodology, followed by a machine learning algorithm-based approach that utilized Natural Language Processing and ML Analysis to categorize comments/tweets as “Positive,” “Negative,” or “Neutral.”
We trained the ML model on over a million records and provided the admin with the ability to manage users, clients, brands, clusters/categories, and sentiment tags.
Additionally, we offered a report component for users to generate and download sentiment analysis detail and summary reports to share with their clients.
Technology Stack:
Frontend: HTML, CSS, JavaScript, jQuery, Bootstrap
Backend: Python 3.7, Django
Machine Learning: Logistics Regression, scikit-learn ML in Python
Database: AWS RDS – PostgreSQL
Server: AWS EC2 – NIGNX
For more such analysis and insights, click here – https://nuagebiz.tech/case-studies/
For more details and personalized assistance, reach out to info@nuagebiz.tech