----"Data is a living thing, it can talk, move and grow."
This repository contains some of my important Data science projects which I developed using extensive machine learning algorithms and interactive visualizations.
Folder:CostAnalysisByCustomerChurnPrediction.
This is my first full stack Data Science project which was developed as a POC by taking a telcom dataset and the model generated approximately 35% profit to the company. I presented this model at Regions Bank (one of fortune 500 companies), they are quite excited and going to test this model on their customer data sample.
I highly recommend viewing this notebook in the below link as github is unable to show tableau and plotly embeds.link: https://nbviewer.jupyter.org/github/kbsriharsha/DataScience/blob/master/CostAnalysisByCustomerChurnPrediction/Churn.ipynb
Folder: FrameworkForSentimentAnalysis-Thesis-I
This is the first part of my thesis. I developed this as a generalised framework for performing social network sentiment analysis on any kind of domain. For case study, I choose Automobile industry for domain, Twitter for social network and Spark as a big data analytics tool.
Note: I highly recommend seeing this model in the link below as github is unable to show interactive visualizations from tableau and plotly.
I highly recommend viewing this notebook in the below link as github is unable to show tableau and plotly embeds.link: https://nbviewer.jupyter.org/github/kbsriharsha/DataScience/blob/master/FrameworkForSentimentAnalysis-Thesis-I/Twitter_Spark_Sentiment.ipynb
Folder:SentimentAnalysis-ActionMail-Thesis-II
This is the second part of my thesis where I performed social network sentiment analysis for a single entity. I used Regions Bank as a case study. In this part, I developed a complete usable Sentiment module which can be used in any program for giving polarity of the sentence just importing this module. This framework uses a action mail system where the system will deliver mails to corresponding entities depending on some preset alarm levels.
All these projects were developed using various machine learning algorithms, statistical analysis and different visualization tools. I kindly request to go through these project notebooks for more understanding.