My professional journey encompasses a diverse array of projects in analytics engineering, where I specialize in solving complex challenges within the business and healthcare insurance domains. From data understanding to Data Ingestion, and from constructing data warehouses to designing data lakes, I excel in crafting star schemas (facts and dimensions) and conducting thorough data analysis via intuitive dashboards.

Moreover, I have spearheaded numerous projects in NLP engineering, delving deep into generation models. For comprehensive insights into my projects and solutions, I invite you to explore the NLP project section.

Furthermore, my expertise extends to various projects in the realm of machine learning, encompassing both supervised and unsupervised problems. Additionally, I have undertaken several ventures in computer vision, demonstrating a multifaceted skill set.

For more details about my projects and each solution, they are described in the every project section.

• Airbnb

  Airbnb's project aims to build an end-to-end data pipeline that extracts Airbnb prices in some of the most popular European cities from Kaggle. Each listing is evaluated for various attributes such as room types, cleanliness and satisfaction ratings, bedrooms, distance from the city center, and more to capture an in-depth understanding of Airbnb prices on both weekdays and weekends.

• Movie Recommendation Website

  This website is dedicated to recommending movies to users based on various recommendation system techniques, including content-based, collaborative filtering, and hybrid approaches. Users can discover new movies tailored to their preferences and interests through these recommendation methods. Additionally, the website provides a search feature where users can input descriptions or details of movies they are interested in, and the system will find relevant recommendations based on the input.

• Sales Forecasting Project

  This project focuses on predicting future sales quantities based on historical data of products. Leveraging real data, which includes detailed information on products and their respective quantities, the goal is to provide the company with accurate sell-out quantity forecasts. This information is crucial for informed decision-making and enables the company to manage its inventory and optimize supply chain processes effectively.

• Watcher

  Watcher is an AI system designed to assist parents in monitoring their children. With the increasing demands of modern life, parents often find themselves preoccupied with work, leaving them concerned about their children's safety, especially when they are home alone. Our project, Watcher, aims to address this concern by creating a system that detects potential dangers and promptly alerts parents. We sought to answer the following questions:

  1. What kinds of dangers could kids face?
  2. How can we detect these dangers in real-time?
  3. How can we make the system function with just a mobile phone and a laptop?

  We have developed a system consisting of two main components: the backend, which receives video streams from a mobile phone camera and analyzes them, and a mobile app that parents use to manage the system and receive alerts.

• Sentiment Analysis App

  A sentiment analysis application constructed using Streamlit (an open-source framework for building web applications in Python). Sentiment analysis stands as one of the most renowned applications of Natural Language Processing, applicable to various forms of data—whether it be text, audio, video, or image. This project's goal is to develop a web application dedicated to analyzing hotel reviews.

• Multi-task Training NER and Sentiment Analysis

  MULTI-TASK TRAINING MODELS WITH HUGGING FACE Arabic TRANSFORMERS: 1- we work on Real Arabic Hotel Reviews to: classify them into good, natural, and bad reviews at the same time, we get the NER for each word. 2- Explore data then Clean it and Analysis the Result. 3- Using Huggingface Arabic Transformers such as Arabert, Marbert, and Qarib. 4- Build and Train a Multi-Task Model then we use it to Predict the test set for every single Task.

• Optimization Accidents hospitals Algorithm with PuLP

  Finding the optimal solution can be divided into five general steps: • Describing the problem. • Gathering the data. • Formulating the mathematical program. • Solving the mathematical program. • Performing some analysis and
  modifying the solution multiple times. • Presenting and analyzing the solution.

• Providing a title for a specific text

  The project aims to generate a suitable title for a given text. We preprocess and segment the data, then train several models. Next, we take a specific piece of news from the test set and find the main topic of the text to provide a suitable title. We extract the most important words from these topics and input them into the previous models to form the title. Then, we measure how closely these titles align with the given text, and titles are generated using the AraGPT2 model.

• Classification of Sounds

  In this project aimed at classifying various sounds (specifically breaking glass and shooting sounds), I utilized the ESC-50 dataset. However, since the ESC-50 dataset does not include shooting sounds, I collected additional sound files, resulting in 532 wave files for analysis.

• Classification Arabic Tweet

  In this project, I aim to classify Arabic tweets into three categories:

  • Positive about the corona vaccine
  • Negative about the corona vaccine
  • Tweets that are unrelated to the corona vaccine

• PDF Vision Project

  In this project, we were tasked with creating a program to read files with the suffix PDF safely and comfortably. The program provides the following features:

  • Reading in eye comfort mode, similar to most smart devices.
  • Ability to convert the background of the pages to a color of the user's choice.
  • Ability to convert the writing to a color of the user's choice.
  • Ability to place a marker on the page of the book to continue reading from it.
  • Ability to move between pages (scaling) depending on the movement of the iris of the eye.
  • Ability to take a screenshot through the eyes.

• Detecting COVID-19 with chest X-Ray

  The objective of this project is to create an image classification model that can predict Chest X-ray scans that belong to one of the three classes: Normal, Virus, and COVID with a reasonably high accuracy using pre-trained VGG-16 & DenseNet models.

• Age Detection

  The Age Detection project utilizes images from "The Indian Actors" dataset. Leveraging Convolutional Neural Network (CNN) models, the project aims to classify the age of actors into two categories: young and old. Using advanced deep learning techniques, the system analyzes facial features and patterns within the images to accurately determine the age group of the actors depicted. Using CNN models, the project achieves precise age classification, contributing to the development of robust image recognition systems.

• Credit Card Clustering

  This project is unsupervised and aims to cluster credit cards based on customer-level data containing 18 behavioral variables, to define an effective marketing strategy. Techniques such as PCA, Kmeans, Agglomerative hierarchy, Gaussian Mixture, and Clustering with PyCaret are utilized for credit card clustering.

• Diamond Price Prediction

  This project is a regression problem that aims to predict the price of diamonds by training multiple models and determining the optimal performance score.

• Music_Genre_Classification

  This project is a classification problem that aims to predict the class by training multiple classification models and determining the optimal performance score.

• Classification of Sounds

  In this project aimed at classifying various sounds (specifically breaking glass and shooting sounds), I utilized the ESC-50 dataset. However, since the ESC-50 dataset does not include shooting sounds, I collected additional sound files, resulting in 532 wave files for analysis.

• Book Recommendation System

  This project consists of the following components:

  1- Popularity Recommendation System:

  This system will recommend books based on their popularity, taking into account user ratings and the name of the author and publisher.

  2- Content-Based Recommendation System:

  Utilizing features such as the name of the book, author, and publisher, this system will recommend books similar to those a user has liked.

  3- Collaborative Filtering System:

  This system will recommend books based on the preferences of similar users. It will analyze user-item interactions to make personalized recommendations.

  4- Hybrid Recommendation System:

  Integrating the results from the content-based and collaborative filtering systems, this hybrid system will provide enhanced recommendations by combining the strengths of both approaches.

• FuzzyClust Image Enhancement

  the task in the project is image enhancement, which is a preprocessing step in digital image processing. Image enhancement techniques aim to improve the appearance or perception of an image, making it suitable for analysis and human visual systems. By enhancing images, we can enhance details, improve contrast, reduce noise, and overall, make the images more visually appealing and informative.

• Water Control System

  The proposed palm tree irrigation control system from Al company is a multifaceted model that combines fuzzy logic technology with the Internet of Things (IoT) and cloud computing techniques. Data is collected from temperature and soil moisture sensors, then the Fuzzy Logic system determines the correct amount of water the pump should supply to the palm trees.

• Ontology NLQ for AI

  This project focuses on developing an algorithm to standardize different ontologies according to the structure provided by the subject teacher. The ontologies consist of multiple XML files, each containing the name of an AI
  algorithm and its relation to it. The primary objective is to enable efficient searching within these ontologies using natural language queries, which are then converted into SPARQL language for querying.