Software Projects
Machine/Deep Learning Projects
An interactive dashboard for predicting Chronic Obstructive Pulmonary Disease (COPD) severity using pre-trained machine learning models. Built with Dash, it allows users to input patient data, choose a model, and visualize prediction results along with ROC and PR curves in real time.
This is a computer vision project where a transformer-based model (Swin Transformer) detects the defects as well as the types of the defect in wafer maps which are produced during semiconductor fabrication process.
Unsafe Selfie Detection
An Image Classification project, where it can distinguish Selfie Images that are taken in an unsafe environment. An Ensemble Model made out of six pretrained models was used for this classification task.
A Multi-class Image Segmentation project where a U-net model is used for the segmentation of both liver and its cancer from a 3D image dataset.
ML Based Wind Turbine MPP Tracker
This project tracks the Maximum Power Point (MPP) of a wind turbine by estimation of wind speed using Semi-Supervised Successive Random Forests.
A Computer Vision project which detects horses galloping in a video. It uses pre-trained Single Shot-Multibox Detector (SSD) model for that purpose.
It is a simple machine learning model which reads from a .tsv file, cleans the text from there using NLTK module, generates a bag-of-word model and after that this model uses an ensemble classifier which tells whether the review is a positive one or negative one for Sentiment Analysis.
It is a Computer Vision project that can detect face and smile. It was created using OpenCV which takes in live feed from webcam. Haar-cascades are classifiers that are used here to detect features (of face and smile in this case) by superimposing predefined patterns over face segments.
Web Scraping Projects
This program uses a web crawler, Scrapy, to explore the PubChem periodic table, grabbing info like symbol, name, and atomic mass for every element. It saves this info in two ways: an SQLite database for easy searching and grouped JSON files so anyone can see how elements in similar families compare. This makes it great for learning about the elements, building cool chemistry tables, and even using the data in research or educational projects.
This Python project scrapes data from Steam's top-selling games, collecting valuable information like titles, descriptions, tags, release dates, review scores, and prices. The project is well-organized folder structure, including subfolders for configuration, utilities, and the virtual environment, suggests a methodical approach to development. The scraped data is saved in a CSV file.
It gets the stock price data of a certain company from API request everyday at a certain time using "pythonanywhere.com", calculates if it went 10% up or down from its previous day closing price and if it does it alerts its user via SMS using "Twilio" about the price movement and a few news headlines from "newsapi" regarding this company so that the users can decide whether they should buy or sell the stock.
This code snippet is a Python script designed for web scraping images from the Unsplash website using the HTTPx library for making HTTP requests and Selectolax for HTML parsing. It includes functions to search for images based on a given keyword, retrieve their URLs, and download the images. The script demonstrates an automated process for scraping images by sending HTTP requests to Unsplash's website and utilizing their API to fetch image URLs, subsequently saving the downloaded images locally based on the search query.
The project is about researching house rent prices that fit some particular criteria for a client on the "Zillow" website. After scraping all the data that is needed from there using BeautifulSoup4 library of python a Google Form is filled automatically per listing using Selenium & Chrome Driver and then which is sent to a Google Spreadsheet.
The Jupyter Notebook demonstrates how to build a simple stock portfolio valuation tool using Python and Google Finance API. It allows users to fetch real-time stock prices and currency information, create a portfolio of stocks with their quantities, calculate the total portfolio value, and display a summary of the portfolio, including the ticker, price, quantity, market value, and percentage allocation for each stock. This tool can be useful for investors to track and manage their stock portfolios.
Miscellaneous Python Projects
It is a simple GUI (Using Tkinter module) Quiz app created using Object Oriented Programming which gathers question data from Open Trivia Database API and track score.
It is '70s famous table tennis themed two player 2D arcade video game. It is an OOP based program and it was created using Turtle library of python.
MIT Scratch & App Inventor Projects
A simple 2d game made using MIT Scratch where player has to move the boat without touching any obstacle or boundary and will get a score point if that passes an obstacle.
A simple coin flip app for Android Operating System made using MIT App Inventor where if a button is pressed it says "Heads" or "Tails" randomly and shows the image of that particular side of the coin.
To find out more projects like these please visit my Github and Repl.it repositories below.
Hardware Projects
This project was designed to protect electrical devices from voltage fluctuations. Here a Single Phase Full Converter, an Arduino UNO, a 4-channel relay module and a custom made multi-tap transformer were used to make this project.
Line Following Robot
Detects black or white line using LDR sensor board (Custom made) connected to Arduino UNO. This project took part in contest ENGINIUS 18’, INNOVENTURE 17’ and TECHFESTA 17’ and won 2nd Runner-up position in INNOVENTURE 17’
This line follower was designed without any microcontroller. Only analog components (e. g. photodiode as sensor, LM385 as comparator, L293D as motor driver etc.) were used for making this project.
Room Temperature Meter
This project was designed to measure room temperature using LM35 analog temperature sensor controlled by microcontroller which gave output to both LCD display unit and 7-segment display.
Simulation Projects
Circuit Design & Simulation
The aim of this project was to ensure both water & energy saving in our houses by turning the pump ON/OFF detecting water level and indicating it. It was a simulation project and used Proteus for the purpose.
IC Design & Simulation
This is a processor with 436 MHz clock frequency, the RTL is synthesized by Cadence Genus and finally the floorplan was designed using Cadence Encounter.
It was an asynchronous counter which could count binary 0-7 or 7-0. At first D-Flip Flops were created with individual schematics and layouts of various gates. Then the whole circuit was created cascading 3 of these D-Flip Flops. Its Schematic & Layout were designed and verified (LVS, DRC) using Virtuoso Cadence software.
Advance Microcontroller BUS Architecture
What basically it does is when the write, select, ready and enable signal is ready it stores data in a given address of memory. It was designed with verilog code using ModelSim software.
