Projects
The MSBA program at Foster is instrumental in giving me the data skills that are needed to support business decisions.In the last 1 year I have used Supervised Machine Learning: Linear Regression, Logistic Regression, Decision Tree, K-Nearest Neighbors (KNN), Support Vector Machine (SVM) - Unsupervised Machine Learning: K-Means Clustering, Dimensionality Reduction: Principal Components Analysis (PCA) - Ensemble Learning: Random Forest, Gradient Boosting - Resampling Method: K-fold Cross-Validation - Shrinkage Methods: Regularization (Ridge/Lasso) - Model Performance Assessment: Confusion Matrix, Precision/Recall, Accuracy, Area under the ROC Curve (AUC)
Hands-on experience in - Python Programming (NumPy, Pandas, Scikit-Learn, Matplotlib, Seaborn) - R Programming (tidyverse, ggplot2, dplyr, plotly, shiny, ggvis, RColorBrewer, rmarkdown, forecast) - Database Management: MySQL, SQL Server, Data Visualization: Tableau, R • Familiar with StatTools,QlikView,Oracle Crystal ball and Google Analytics.
​
Below are a few individual projects that broadened my data skills. click on them to read more
Coursework and Group Projects
Finding signals in data to distinguish leisure, business and bleisure travel for Hilton Hotels using Adobe Analytics
Adobe Customer Journey Analytics
Objective: To find signals in dataset provided by Hilton and distinguish whether a trip represents business, leisure, or “bleisure” travel, and make recommendations on digital properties
​
Dataset: Hilton dataset from January 2021 to Oct 2022
​
Approach: I analysed the data based on number of sessions for a booking,time spent/session, membership tiers, devices used, number of children booked, number of nights booked. Further dived deep into number of children=0 to understand business or bleisure travellers
​
Findings: Indicator1: Number of children travelling equals 0 means business or bleisure, to find purely leisure Indicator 2: Time spent/session less for Business destinations, events/reservation more for leisure travellers
Tariff Barrier Analysis of Automobile Export from India
Excel, World Trade Organisation Tariff Analysis, TradeMap, WITS database
Objective: To find the tariff and non-tariff barriers that hinder trade of automobiles from India
​
Dataset: Monthly, quarterly and yearly trade flows of 10000 products across 220 countries
​
Challenges: Combining metrics such as volume of exports, world rankings, average tariffs,Trade Balance,etc. across 3 different databases and weighing tradeoffs to finalize top 3 potential destinations of export.
​
Findings: Argentina, Phillipines, Mexico and Australia as top export or joint venture destination and recommended four changes in govt policy to enjoy unutilised tariff concessions