LARAVEL ECOMMERCE WEBSITE
Link to the project code
Web Developer | Digital Marketing Executive
Expert Web Developer with over 6 years of experience in designing and optimizing high-performance websites within an agency setting. Possesses wide knowledge across various programming languages and frameworks, with advanced proficiency in both front-end and back-end technologies. Skilled in applying best practices for web performance, including A/B testing and analytics for data-driven improvements. Experienced in utilizing cloud platforms and containerization technologies to enhance development workflows and application scalability. Demonstrates expertise in integrating digital marketing strategies to enhance online visibility and engagement. Proven ability to leverage agency experience to deliver robust, high-quality web solutions.
Ecommerce Website
CASE STUDIES (DATA PROCESSING, MARKETING DATA ANALYTICS & MACHINE LEARNING)
The Notebook Contains:
. Data exploration for initial processing . Handle missing values using statistical tests . Compare 3 methods to detect and remove outliers based on data distributions
. Compare 3 feature selection methods
. Handle imbalanced classes
. Apply ML algorithms to predicts customers' response to the compaign using cross validation
. Try different data copies obtained from previous phases (standardization, outliers detection and handling imbalanced classes)
. Plots, comments and texts to explain and clarify the code and the results
SQL
Maven Fuzzy Factory Dataset (Imported to MySQL Database)
Link to the queries and the created CSV files
The queries created 18 Views Exported to CSV Files to be used in Tableau, and answer these questions:
. What is the conversion rate for each utm source, utm compaign and used device (total, per year_month or per year_quarter)?
. How much is the company revenue per year_quarter?
. What are the most sold products (ordered by revenue)?
. What are the most returned products?
. What are the most visited pages?
. Do users tend to buy at their first or later visits?
. Does the company new users increased over the years?
. How many people make it to the cart page on the path to buy MrFuzzy product? (aconversation funnel analysis between the product page and the cart page)
. How many days between consecutive orders? (to cluster active customers)
. How many users returned orders once, twice or never ask for refund?
. Who are the top 20 buyers and what are the utm sources they use?
. What are the products that are sold together?
. What is the trend in brand and nonbrand searches over the years?
PYTHON (CNN)
build a multi_task (one) model to predict (age, gender and race) using UTKFace (20k+ cropped face images with age, gender and ethnicity)
Link to jupyter notebook file
. The code contains transfer learning, image generating, augmentation and one model with three branches for age, race and gender
. The model is a prototype trained for 20 epochs and needs to be tuned.
FLASK & ML
Scrap movies titles from IMDB according to detected emotion in realtime.
Link to project folder
a flask app that uses a trained machine learning model to detect the seven basic emotions in realtime using webcam then scrap the top rated movies of a specific genre according to the detected feeling
PYTHON (CLASSIFICATION)
A jupyter notebook to classify stroke stage
Link to project folder (with dataset)
build and train a model using heart disease dataset on kaggle (contains patients history with five classes that indicate the absence(0) or presence of heart disease in four stages(1,2,3,4))
the model uses SMOTE technique (oversampling) to solve the "class imbalance" problem in the dataset. This technique increased the model accuracy from 60% to more than 95%
TIME SERIES (MATLAB)
predict the next 4 steps of a user in a MANET network to be used later to improve MANET protocols using NARXNET neural network
Link to the project
proposed training model
using ELM autoencoder for NARNET weight initialization
compare the proposed model (red line) with 2 other common algorithms
