Bazarbay Alisher
A highly focused individual with strong communication skills and solid science background, passionate about data science, machine learning and FinTech. 3rd year student, majoring in Data Science | City University of Hong Kong.
Investing
The notebook downloads the historical data for desired assets(stocks, crypto, etc). It calculates and plots the bollinger bands and ichimoku cloud indicators. Also, it optimizes the portfolio of desired assets (Markowitz theory and Sharpe Ratio metrics), calculates return and volatility, plots returns/correlation matrix.
Page Rank
Simulated a random walk and calculated the L1 distance between the normalized degree vector and the empirical frequency vector; Calculated the PageRank of the graph (PowerIteration and Monte Carlo methods);
Data Transfer
Data transfer from MongoDB to AWS S3 bucket for further data visualization in AWS QuickSight. Data is extracted by Lambda function (with pymongo library) configured in a VPC with a private subnet. Further, collections are uploaded to S3 (boto3) with a VPC gateway endpoint. The route table is shown on the scheme. The database and handler are connected via peering connection (voiding connection via global network).
Instance Search
I implemented an application that can perform instance search using SIFT and Color Histogram methods. OpenCV library was used with common SIFT algorithms family interface. Several scale invariant algorithms were tested. I also tried to combine the SIFT and BRISK descriptors to solve the problem.
Virtual Drawing
Application to detect the object based on specific color range; it detects the contours, pins the point and draws. In addition, you can fix the points and connect them to form a shape.
Maze Generation
Maze Generation in Python. In this application, I used Recursive Backtracker algorithm to generate a maze, with usage of Pygame library.
Fraud Detection
Analysis on a real-world dataset on credit card fraud detection. Data Exploration, Vizualisation, Dimension reduction, Classification, Model Evaluation. Trained 3 Classification models: SVM, Random Forest Classifier, AdaBoost Classifier.
Online Shop Website
This is my high school project. I wrote the computer hardware store website for my client. It has all essential functionalities except payment: registration, sessions, database to store the users and products, cart, sorting, searching, categories and admin panel.