PROJECTS

https://www.sanildoshi.com/projects

YES!

https://www.sanildoshi.com/projects/LakersIn4

“Is LeBron the GOAT?” - A Deep Dive into NBA Statistics

www.islebronthegoat.xyz

The project serves as a data-driven playground for basketball aficionados and casual fans alike to explore the question of LeBron James' greatness.   

While it provides significant insights, the debate around the Greatest Of All Time is far from settled. Future iterations of the project could include real-time data updates, inclusion of additional players for comparison, and even an AI-powered chatbot to guide users through the debate.

To create a comprehensive narrative, a multitude of data sources were identified and scraped using Python libraries. These sources included player statistics, career milestones, and other metrics that contribute to the GOAT discussion.

Understanding data can be complex. To bridge the gap between numbers and comprehension, the front-end of the website was built using React.js and Vite. This tech stack allowed for the creation of interactive visualizations and optimized assets for production.

The website was deployed end-to-end using a suite of AWS services:
‍
- Amazon S3: For storing static resources
- Route 53: For DNS management
- CloudFront: For CDN and efficient content delivery

These services ensured that the website remained highly available and easily accessible to a global audience.

In closing, the "Is LeBron the GOAT?" project serves as an innovative intersection of data science, front-end design, and cloud computing to quantitatively explore one of basketball's most heated debates. By leveraging an array of technologies — from Python for data collection and analysis, to React.js for user-friendly interactive visualizations, all hosted seamlessly on AWS — the project offers a comprehensive, data-driven platform for both seasoned analysts and casual fans.

Maybe one day we will know the answer to the age-old question of LeBron's greatness, but for now, let's cherish the opportunity to watch this living legend on the court before he hangs up his jersey for good.

https://www.sanildoshi.com/projects/stat-arb

"Quantitative Trading Algorithm"

A look into Statistical Arbitrage and Pairs Trading Strategy in a Financial Market

GitHub Repo

In today's volatile financial markets, trading strategies need to be both robust and dynamic. With that in mind, I engineered a Quantitative Trading Algorithm focused on Statistical Arbitrage and Pairs Trading Strategy. This ongoing project seeks to automate trading decisions based on statistical measures, risk assessments, and real-time data analytics.

Programming Languages: Python, SQL
Data Sources: Yahoo Finance API
Database: PostgreSQL
Data Analysis Libraries: Pandas, NumPy, Matplotlib, Seaborn, Scikit-Learn
Development: Jupyter, Git

Armed with a wealth of data, we ventured into Exploratory Data Analysis. Utilizing Pearson correlation coefficients, we isolated asset pairs that showed promising correlations. This is pivotal for any pairs trading strategy, as the essence of the approach is to go long on one asset while shorting its pair when divergence occurs.

To lend statistical power to our strategy, we conducted the Augmented Dickey-Fuller test to confirm cointegration between asset pairs. Cointegration ensures that the pairs will converge over time, even if they diverge in the short term. We also utilized the Hurst exponent to ascertain the mean-reverting nature of the asset pairs. Lastly, Monte Carlo simulations along with Beta and Sharpe Ratio calculations were performed to assess the risk and potential returns.

Beta: Measures volatility of an asset in relation to the market
Sharpe Ratio: Indicates the risk-adjusted return of the asset
Hurst Exponent: Reveals whether the asset pairs is mean-reverting, trending, or a random walk
Monte Carlo Simulation: Method to predict possible outcomes of an event through repeated random variables

The backend scripts are hosted on AWS Lambda and API Gateway, allowing us to execute code without provisioning or managing servers. For the front-end, React.jsx was utilized, and the entire architecture was hosted using AWS services like S3 Buckets, Route 53, and CloudFront to ensure high availability and efficient content delivery.

In the volatile world of trading, where every millisecond counts, our Quantitative Trading Algorithm stands as a testament to the power of data analytics and computational excellence. As we continue to fine-tune our algorithm and adapt to new market conditions, we remain committed to providing a sophisticated, risk-adjusted trading strategy that thrives in complexity.

https://www.sanildoshi.com/projects/dedropz/shoutoutbanshee

"DeDropz"

A Web 3.0 E-commerce platform on the Solana Blockchain

www.dedropz.xyz

Technologies Used: AWS (S3, EC2, Route 53, CloudFront), React, HTML5, CSS3, hel.io
‍
‍Duration: January 2023 to May 2023

In the burgeoning landscape of non-fungible tokens (NFTS), DeDropz emerges as a specialized e-commerce platform that caters to a niche, high-valued NFT community: DeGods. The platform is designed to offer exclusive, branded collectibles, all while leveraging the speed and security of the Solana blockchain.

DeDropz is built on the Solana blockchain, chosen for its high throughput and low latency, which ensures seamless and efficient transactions.

Payments on the platform are facilitated through hel.io, an external web3 payment platform. This allows users to make seamless transactions using cryptocurrency, enhancing the overall user experience.