vshah5,Valay Shah,github

car-auction-regression-simulation-optimization

A full statistical methodology is used to determine the bid price of cars in an upcoming auction to help a dealership maximize their profits and stay within their budget. Data was first cleaned and prepared for regression analysis, then the car auction was simulated to determine the bid percentage which maximizes the chances of winning the auction for a handful of cars, and lastly optimization was used to determine which cars to bid on according to the budget and available floor space of the dealership.

churned-customers-naivebayes

Using the data provided for the past 3 months, I have created a model by classifying the first two months as a customer’s happy phase and the 3rd month as a customer entering the sore phase. Based on this, I created a model which predicts which customers will enter the churn phase next month. By using this model, we will be able to identify and prevent customers from churning. Additionally, we can determine what the appropriate discount given to unsatisfied customers should be to retain them for a longer period of time and maximize our profits. In simple terms, the methodology to build this model utilized two statistical methods, logistic regression as well as Naïve Bayes. Both methods were used in order to determine which one of them gave a more powerful and accurate predictive model. Many different variables were inputted into the model such as roaming minutes, network minutes, amount of data used, and how long the customer had been with us, among others. Next, I was able to discern which variables had the highest predictive capacity for customers churning. After running both logistic regression and naïve bayes techniques, I found logistic regression to produce a model which produced 93% accuracy in predicting the churn of customers. Combining this model with historical information on how discount percentages led to a certain percentage of churning customers retained enabled me to produce a table which identified what discount percentage we should offer to our sore phase customers in order to retain them.

database-creation-and-querying-case

A database was created based on the case info, data fields were inserted into the database, and finally SQL queries were used to do the following: a) Provide an alphabetical list showing common and legal names of all Ontario charities. b) Generate a list with employee first name, last name, and number of articles published. c) How many times was each article viewed, from most to least popular? d) List the common name of each charity, along with its sector and sum of all donations made during 2020, starting with the highest donation total.

equity-research-report-and-stock-pitch

This repository contains the equity research report (pdf), stock pitch slide deck (ppt), financial models (xl), and a couple other preliminary analysis slides all conducted for the purpose of researching Mead Johnson Nutrition company. All the documents were prepared by me for a potential investment opportunity when I was a Senior Analyst for the Student Investment Fund at the School of Accounting and Finance at the University of Waterloo.

federal-database-visualization-and-analysis

Using a database of over 7,800 enterprises, the Survey of Innovation and Business Strategy was a joint project undertaken by Industry Canada, Department of Foreign Affairs, Trade and Development, and Statistics Canada. I used the extensive database to gather insights and intelligence about the food and beverage processing sector in Canada and how this sector performed compared to the manufacturing sector as a whole.

general-motors-ventilator-production

The purpose of this report is to provide a framework for crafting a distribution strategy for GM’s Ventilator Project for U.S states. For this, two optimization models were used, one which maximized the number of ventilators distributed and the second model which minimized the cost of distribution.

hack-the-case-scotiabank-sas-clustering

Utilized SAS, SQL, and SAS clustering to clean SCENE and Scotiarewards customer database, and conduct clustering analysis to derive insights on Scotia customers using credit cards. Used the clustering analysis to derive unique solutions to help Scotiabank retain and attract new customers towards their credit cards. Presented our findings to a panel of judges comprising of Deloitte, SAS, and Scotiabank staff; finished as finalists in the Hack the Case competition.

hierarchical-clustering

Cereals.csv contains information on 77 different cereals and their characteristics. For the analysis, all cereals with missing values were removed, hierarchical clustering was applied to the data using Euclidean distance, and normalised measurements. Finally, the dendograms for single and complete linkages were compared and a recommendation for the number of clusters and cluster membership was determined.

java-calculator

Using Java and Apache NetBeans, I created a mathematical calculator which also has a dropdown help menu for remainder, sin, cosine, and tan buttons that take the user to the Wikipedia entries of those concepts.

naive-bayes-and-confusion-matrix-on-bank-customers

Naive Bayes, Confusion Matrix, and ROC Analysis were conducted using R to determine how different variables lead to a customer of a bank taking out a personal bank loan.

preventing-recidivism-in-state-prisons

Using R, logistic regression was performed to evaluate the prisoner reform program based on the data of state prisoners. From the total 432 convicts which were released from state prisons, 50% of them were chosen at random to receive financial aid. Prisoners were then tracked, and additional metrics were observed and their impacts on prisoners getting re-arrested were measured in this analysis.

python-tax-calculator

This program asks for a price, calculates the price with tax #and outputs the result.

python_name_info

This program asks for the user's first and last name and outputs the number of letters in the full name as well as what letters it begins and ends with along with what the initials are.

quesada-simluation-and-optimization-case

This case was created by my team as part of the Art of Modelling final group project. The case utilizes the simulation and optimization methodologies in order to solve the problem of long queue times at a Quesada franchise. The Case Problem pdf describes the problem at Quesada and the Simulation and Optimization pdf is a teaching note formatted sample solution for those trying to solve the problem using simulation and optimization. The excel doc is our solution of the problem.

sentiment-analysis-tweets

In literal terms, sentiment analysis is an analysis of people’s sentiments. It is a technique that classifies text data scraped from the internet based on the predicted underlying sentiments. An election campaign team can use sentiment analysis to identify individuals to target, who are more likely to vote for their candidate based on their online sentiments. Sentiment analysis can be used on its own to classify and interpret text-based data or used as a preliminary step for further analysis with logistic regression, naïve bayes or support vector machine.

simulating-theater-experience-simul8

Created multiple simulation models on SIMUL8 software for a theater company to modify number of staff members, starting times, kiosks, and guest arrivals in order to maximize the number of customers as well as maximize profits for the theater.

sql-database-queries-for-pizzeria

Created entire database from scratch for a local pizzeria. Generated data, created a barker-notation model, created the database using phpmyadmin, then answered important managerial questions through sql queries. Lastly, presented all the information with answers to the queries which will help the business sustain and grow itself. Presentation component was ranked the highest in the class with 97% grade.

sql-yelp-queries-predictiveanalysis

Derived insights from a massive Yelp database regarding companies, reviews, and users. Created a methodology for predictive analysis to see which companies will survive in future years.

student-demographics-impact-on-academic-performance

This group project observed the impact of different student demographic variables such as parental education level, lunch option, ethnicity, gender and others on student's academic performance. The tools utilized were regression and conditional probability. R and Excel were used for these methodologies.

survival-analysis-of-companies

The purpose of this project is to present two models, one with logistic regression to see how certain financial variables impact a company’s probability of default, and another model to predict the expected time to bankruptcy for any given company.

unb-data-challenge-tableau

Created a visually appealing infographic and aesthetically pleasing dashboard to present insights about crime in Nunavut in accordance with the Canadian government's mandate to achieving their sustainable goal of Peace, Justice, and strong institutions. Rewarded with 1st place in the presentation round and 3rd place in the infographic round by a diverse set of professional judges.

vba-automated-checklist

This is a skeleton version of an automated VBA checklist I created for my prior workplace to help our internal auditor with internal control requirements. This checklist was manual prior to the creation of this automated checklist and it helped save time, save paper, was user-friendly, and added additional features which enforced segregation of duties. The checklist functioned by the preparer first going into the file and filling out their checkmarks/NA for each fund preparation requirement and then simply pressing a button to lock their column of the spreadsheet and have their initials and timestamp recorded and locked. This process was repeated by the reviewer and final manager. This skeletal version is trimmed down from the original document and the code has been locked for proprietary reasons.

vshah5 Goto Github PK

Valay Shah's Projects

Recommend Projects

Recommend Topics

Recommend Org