Python, Machine Learning, Geoinformatics, Feature Extraction

Shardul Rane

********@*******.*** +1-716-***-**** https://www.linkedin.com/in/shardul-rane-06142b115

EDUCATION

University at Buffalo, State University of New York May 2021

Masters of Science in Data Science

University of Mumbai May 2017

Bachelor of Engineering, Computer Engineering GPA: 7.52/10.0

Relevant Coursework: Statistical Data Mining, Machine Learning, Linear Algebra, Deep Learning, Data Intensive Computing, Probability, Data Modelling and Query Language, Data Visualization

TECHNICAL SKILLS

Programming Languages: Python, R, JavaScript, MATLAB, Java, C++, HTML5, PySpark, Jupyter

Technologies and Skills: TensorFlow, Keras, Pandas, Scikit-learn, Numpy, SQL, SQLite3, ETL, BigQuery, REST API, Spark, Hadoop, Azure, AWS EC2, MXNet, GLM, Git Version Control, PowerBI, Tableau, Exploratory Data Analysis, Anomaly Detection, Natural Language Processing, Topic Modelling, Predictive Analytics, Adobe Analytics, Google Analytics, Excel, PowerPoint

ACADEMIC PROJECTS

Anomaly Detection Using Unsupervised Techniques University at Buffalo [Link]

oTraced Anomalies for varying threshold with One-Class SVM, Stata ARIMA, Auto encoder Neural Networks LSTM, K-means and Principal Component Analysis on ‘Expedia hotel reviews’ and ‘Pressure sensor’ time-series Kaggle dataset.

oTools Used : Scikit-learn, Matplotlib, Keras, pyod, plotly, statsmodel

Classification Model Testing to Predict Soccer Game Outcome University at Buffalo [Link]

oImplemented Ensemble Learning with 71% accuracy for Spatial Clustering KNN, Classification Algorithms RandomForest and Support Vector Machines on multiclass dataset (20 years) to predict game outcome

oData resampling methods used : 10 Fold Cross-Validation, Bootstrapping and Factor Analysis of Mixed Data

Examined Effect of ‘US China Trade War’ on Domestic and International Companies University at Buffalo [Link]

oInvestigated day wise effect of trade war tariffs in sectors of Electronics, Textile, Distribution services and Energy companies by Stock data analysis. Tools Used : Yahoo Stock API, Beautifulsoup, Matplotlib, CandleStickPlot

Music Industry Pattern Study based on Billboard top 100 Songs University at Buffalo

oAnalyzed the trends in music across 7 Genres (60 years) based on 15 audio features from Spotify API with Seaborn, Pandas

oImplemented Latent Dirichlet Allocation topic modelling in Spark to cluster and predict topics for given song lyric

Open Source Contextual Notes Application Based on Linux Platform University of Mumbai [Link]

oDeveloped software with Linux X-Server to identify background application and to display notes affiliated to it

oTools Used : X-Lib, Linux Shell-scripting, Python

Twitter Sentiment Analysis University of Mumbai [Link]

oAnalysis of tweets based on sentiment & polarity with Twitter API using Python, BeautifulSoup, Textblob, Tkinter

PROFESSIONAL EXPERIENCE

Performics, Publicis Media - Mumbai, India December 2017 – June 2019

Web Analyst, Data Analytics

Developed ‘User Behavior and Intent Prediction’ Proprietary Tool

oBased on user’s search query and historical data, bucketed users in 8 different categories with accuracy of 84% in Python, by estimating purchase intent. Methods : K-Means Clustering, Principal Component Analysis, Random Forest Classification

Implemented Custom Revenue Attribution Model for E-commerce Websites

oBuilt an ETL based attribution model based on user behavior and updated First touch attribution in E-commerce journey to assist in planning marketing strategies for 2018 (Q2, Q3). Tools : Python, BigQuery, JavaScript, Google Analytics, App Engine

Designed Data Acquisition Framework for Content Performance Measurement of Bloombergquint Website

oCreated framework with JavaScript and REST API to measure user level article performance in Adobe/Google Analytics

oBased on article data analysis website traffic increased by 14% during Indian Election (2019)

Built Automated Solution for Server-side Placement of User Tracking Codes for DBS Singapore, Nestle and SUTD

oReduced time complexity of 10,000 remarketing codes from 9 months to 20 hours. Tools : Python, Google REST API, JavaScript

Implemented Website Revenue, Product Performance Analysis Visualizations and Dashboarding, A/B Testing & Audits

oDomestic clients: ThomasCook India, Reliance General Insurance and International clients: Securitas (Sweden)

CERTIFICATIONS AND COURSES

Neural Networks and Deep Learning Certification (Coursera), Google Analytics Advanced Certification, Java Advanced Certification (CMIT), Apache PySpark (LinkedIn), Amazon Web Services Machine Learning Essential, Tableau Certification (Udemy)

Contact this candidate