Download PDF

Profile Overview

  • Data Science consultant with a passion for turning data into products, actionable insights, and meaningful stories.
  • 3 years of relevant experience in Data analytics.
  • Strong background in applied machine learning, data science, sports analysis, statistical text mining.
  • Experience in setting up Data Science practice from scratch.
  • Internet of Things expert with extensive experience working with  Bluetooth low energy technology, sensors like accelerometer, proximity, light, step detector. 
  • Proficient in R, Python, Tableau, Minitab, SQL, NoSql DB and advance excel functions
  • Ability to learn and apply appropriate analytical tools and techniques, develop custom algorithms to accomplish business objectives.
  • R trainer and speaker. 

Work History

Education

20122014

Masters in Computer Application

20032005

Bachelor in Information Technology 

Product classification- Kohls

Kohl's, is an american department store retail chain into E-commerce domain. I created a model for product classification. 

Data analysis story

  • Process
    • Modelling --> Model --> Evaluation -->Prediction
  • Model Evaluation Metrices
    • Accuracy = # of correct prediction/Total # of records
    • Confusion Matrix - True positive rate, F-score
  • Data mining techniques
    • Baseline model
    • Text Processing- Cleaning irregularities, punctuation, removing stop words, stemming
    • Decision Tree
      • Entropy calculation
    • Random Forest - Ensemble approach
    • Support vector machine -Large margin classifier

IOT Project - Kastle Presence

KastlePresence from Kastle Systems is a Smart office platform built on the latest in mobile technology. Cutting-edge features including hands-free access capabilities, location-based services, personalized security, and powerful occupancy data deliver the Internet of Things to your workplace.

  • Hands-free access
  • Occupancy insights
  • Security for the individual
  • Location based services and amenities

Technology

  • BLE- Gatt profile.
  • iBeacon and Eddystone profiles.
  • Redis and Mongo Db for big data.
  • Sensors like accelerometer, light , step detector and  proximity.
  • Redis and MondoDB for BIG and Unstructured data.

Data science

  • Predicting calibration value for a user based on device characteristics
  • Analyzing sensors data to calibrate for new devices
  • Analyzing sensors data to detect failures 
  • Predicting occupancy 

Recommendation Engine - Reliance Games

Currently working on building a recommendation engine for in-app purchases for a specific game. The engine should be able to predict which users to show a recommendation.

Data analysis story so far

  • Process
    • Exploratory --> Data cleaning --> Dimension reduction --> Clustering --> Collaborative filtering