Joseph Zahar's Projects
This analysis focuses on the optimisation of a car tow hook by evaluating von Mises stress and deflection values using the Finite Element Analysis (FEA) using a computer software (Abaqus) which uses the Finite Element Method (FEM) simulating the hook is under load. The results were plotted using Python.
The unique mechanical properties of the articular cartilage depend on the interactions between the chondrocytes and the Extracellular Matrix (ECM) that maintain the tissue. the main aim of this experiment is to determine the optimal in-vivo environment for chondrocytes to allow greater cell viability and gag concentration.
Investigation of a simulation of laminar incompressible flow through a stenosis of the aorta using varying orders of discretisation. From the data obtained, the fine, medium and coarse mesh types were compared graphically and qualitatively.
The following repository make use of the Holt Winter and ARIMA models to forecast the demand of one specific food ingredient for a fast food chain over 6 months based on historical data.
This repository presents an interactive dashboard that allows exploring American Sign Language (ASL) landmarks in 2D space. By selecting different ASL categories and IDs, users can filter the displayed landmarks and study the hand shapes and movements associated with specific signs.
Given a dataset of customer data, containing 200 samples and four features: gender, age, income and score, we will build a K-Means clustering algorithm and fine tune its hyper parameters in order to segment these customers into different groups
An intelligent energy platform company called Kaluza could possibly help the Orkney Islands solve this major inefficiency. Rolling out Kaluza’s system of smart heaters which control the heating of storage heaters as well as hot water cylinders could potentially allow households to use the energy that is currently being curtailed. Hence, investing in Kaluza’s smart heaters could be of financial interest for all parties involved; the consumers, the government, as well as the wind energy generators. If this were to be achieved, consumers would have access to cheaper energy, the local government could avoid the cost of expanding the 40MW cables to handle more energy, and the wind energy generators could expand their revenue stream.
Given data of student loan applications, we are going to build a decision tree and a random forest classifier to predict defaults in students loans. We will also provide a strategy to tune the hyper parameters of our models.
We will be exploring a type of batch processing called Extract, Transform and Load (ETL). Extract, Transform and Load or (ETL) does exactly what the name implies. It is the process of extracting large amounts of data from multiple sources and formats and transforming it into one specific format before loading it into a database or target file.
The goal of this competition is to classify isolated American Sign Language (ASL) signs. We create a TensorFlow Lite model trained on labeled landmark data extracted using the MediaPipe Holistic Solution to improve the ability of PopSign to help relatives of deaf children learn basic signs and communicate better with their loved ones.
In the project, I undertake the tasks of collecting data from multiple sources, performing exploratory data analysis, data wrangling and preparation, statistical analysis and mining the data, creating charts and plots to visualize data, and building an interactive dashboard. The project will culminate with a presentation of a data analysis report, with an executive summary for the various stakeholders in the organization.
We will apply foundational Python skills by implementing different techniques to collect and work with data. Playing the role of a Data Engineer and extract data from multiple file formats, transform it into specific datatypes, and then load it into a single source for analysis. We will also implement webscraping and extracting data with APIs to gather informations from new data sources. This project aims to collect large datasets from multiple sources and transform them into one primary source, and web scraping to gain valuable business insights all with the use of Python.
In this study, we investigated the effect of storage conditions (4°C & -20°C) on the mechanical properties of fascicles extracted from a single male Wistar rat's tail swathed in phosphate-buffered saline solution (PBS). Tissue stress and viscoelastic properties were recorded and analysed for both failure and stress relaxation test.
Using the first 800 instances in the file “winequality-white.csv”, we will build a k-Nearest Neighbours classifier in Python to predict wine quality depending on eleven different factors.
The function provided act according to the input provided and aims to replace the largest of both elements.
Comparison of the mechanical properties of different prosthetic liner's materials from the study conducted by Cagle, et al. in 2018.
Mechanical testing of different specimens with different size, more specifically Stress Relaxation tests for small and large specimens
Starting in October 2012, Medicare began penalizing hospitals who had excess readmissions for each of the six readmission measures (AMI, CABG, COPD, HF, HIP-KNEE replacement, PN) reported in the Hospital Readmission Reduction Program data.
This repository is dedicated to the exploration and analysis of machine learning-labeled surgical instrument annotations. Through the development of a comprehensive dashboard, we aim to uncover the underlying patterns and trends in instrument annotations for various surgical procedures.
Modeling of Metastatic Cancer and its Treatment using Swarm Intelligence Based-Control. This study investigates the use of swarm robotics and nanotechnology to target cancer cells autonomously and efficiently, focusing on exploring new control algorithms inspired by the combined behaviour of social animals.
The aim of this repository is to webscrape flight prices data from different airlines and feed them directly into an interactive dashboard built using dash and Plotly.
An analysis of muscle activation, when increasing the repetitions of a single-arm dumbbell curl, supported by raw data of muscular activity obtained from sensors and processed using MATLAB
Using the Fashion-MNIST data, we will build a random forest classifier as a naïve benchmark to compare the efficiency and accuracy of a fully-connected neural network. We will also explore the hyper-parameter tuning of our neural network by manipulating the size of the different types of layers, the loss function, the epochs and the batch size located in the initializing phase of the model.
Various fracture fixation methods have been utilized to promote fracture healing, based on the type and location of the fracture. Intramedullary (IM) nails have been long used to treat spiral fractures located distal to the knee and ankle. Consequently, this analysis aims to propose a novel IM nail design for the treatment of spiral fractures based on a given case (33-year-old man), which aims to reduce stress shielding and avoid re-surgery.
The algorithm we present allows users to track objects in real-time by either uploading a video or accessing their computer camera and manually selecting the object of interest. The movements of the object are then monitored, providing precise and valuable information with numerous applications, from security and surveillance to robotics and beyond
This repository includes four different chunks of codes that aims to optimise the supply chain distribution of a virtual a company to minimise costs while maximising revenues. The virtual company built a factory and one warehouse in one of the five regions it is serving. The goal here is to find both the long term decisions that includes where to build other factories and warehouses as well as short term decisions like monthly ROP and POQ.