## [Follow me on Github](https://github.com/fagrahmed12)



Diamond Price Prediction


EDA project about data scientist salaries. It's apart from my training in [SHAI For AI | شاي للذكاء الاصطناعي](https://www.linkedin.com/company/shaiclub/) B-2022 
The data contains information about several companies. This information shows the size and location of each company, where the employees are located during the employment period, the areas of work in each, the type of employment if it is full-time, part-time, contract, and freelance, the level of experience, the total salary paid, the currency in which Pay it and its equivalent in USD and the rate of completion of the work

Project Tools: This data was analyzed using Power BI and Python.

Some insights from the data:
1-The demand for data scientist roles has been increasing over the years.
2-The senior level is the most frequency
3-Most job a full time
4-Most jobs are fully remotely
Notebook: [shorturl.at/frTW8](http://shorturl.at/frTW8)





Data Science Job Salaries

This project is designed to  Create Automated-ML-Application-for-Exploratory-Data-Analysis-Streamlit-Deployment with Python, streamlit.

Highly comprehensive analysis with all  data cleaning, exploration, and visualization has been in the back end with Python and its intention to create it like a NO CODE PLATFORM

https://github.com/fagrahmed12/Automated-\_EDA


Sunset fishing is a beautiful time of day to relax and get in touch with nature

Automated EDA

Heart disease is very dangerous for people of any age. Lot's of people are dying every year due to heart disease. If someone with heart disease follows proper rules and regulation he might live ok. So the important part is to determine if someone has heart disease or not. In this notebook, we are going to do exactly this.

### Steps

1\. Understanding the problem


2\. EDA


3\. Data preprocessing


4\. Modeling


5\. Evaluation

https://github.com/fagrahmed12/Heart-Disease-Prediction


Heart Disease Prediction

This project will be based on a dataset obtained from the UCI Repository. The dataset consists of 1030 observations under 9 attributes. The attributes consist of 8 quantitative inputs and 1 quantitative output. The dataset does not contain any missing values. The dataset is focused on the compressive strength of concrete. The attributes include factors that affect concrete strength such as cement, water, aggregate (coarse and fine), and fly ash, etc… The objective of this project is trying to predict the concrete compressive strength based on important predictors. The study will consist of evaluating the impact of different factors such as cement, water, age, fly ash, and or additives. We will evaluate the components that are highly correlated with concrete compressive strength and other components that are less influential and can be neglected through visualization or correlation matrix. In this study, we will use different machine-learning techniques to predict concrete compressive strength. Different modeling techniques will be used for the prediction. The modeling technique will include multiple linear regression, decision trees, and random forests, etc. A comparative analysis will be performed to identify the best model for our prediction in terms of accuracy. The best model will be helpful for civil engineers in choosing the appropriate concrete for bridges, and house construction.



https://github.com/fagrahmed12/Concrete-Compressive-Strength





Concrete Compressive Strength

The aim of this Repository is to be able to recognize text from an image file using the e ArabicOcr Library in the Python Programming Language.

e ArabicOcr is an Open Source library for Optical Character Recognition(OCR). We will be using e ArabicOcr to print the recognized Arabic text given an input image.

https://github.com/fagrahmed12/OCR_AI_Team

How to Calculate OCR Accuracy
1\. The Quality of Original Source Images
2\. The Quality of OCR Engine

These factors lead to improving the quality of reading and making the words
appear as correctly as possible:
 Good Quality of Source Images
Before using OCR, make sure you can read the images with your own eyes. If you, with
your own eyes, can’t see the image clearly, make sure the original source images are not
damaged AND wrinkle-free. So, use the cleanest and most original files for better results.
 Right Size of Images
OCR engine needs to read source images not only the ones with the best quality but also
the right resolution. Make sure the image is resized to the correct size, which is usually
about 1 / 10 of the original size (1.5 mm x 1 mm) or less. This way, the result will be
more accurate.
 Remove Noise / Denoise
Human eyes can’t even read documents that have many noises, so does the OCR engine.
Noises make the engine difficult to read original sources and it can decrease the OCR
accuracy. If the image has background or foreground noise, remove it to get a higher
quality data extraction.
 Increase Image Contrast
How do you see white papers with light grey ink? You -and the OCR engine must be
uncomfortable reading such papers. Thus, try to increase the contrast between text and
background brings more clarity to the output. The best contrast will help the OCR engine
to read images accurately.
 De-skew Original Source
No one wants to read papers upside down. Thus, make sure you get the image in the right
format and shape (text should appear horizontal and not inclined). The image can be
rotated by tilting it to one side, turning it clockwise or counter-clockwise, and turning it
back to the other side.

I have also used easyocr to convert any language in the image to text
using OCR techniques
This was their output
https://github.com/fagrahmed12/OCR_AI_Team/tree/master/english
https://github.com/fagrahmed12/OCR_AI_Team/tree/master/french


In the case of Arabic, it was the same result as Arabic that appeared in arabicocr
previously
Here was the knowledge that the quality of the image affects the quality of OCR
Changing the package does not affect the quality of the reading


Optical Character Recognition

This project focuses on pandas library usage and simple statistics methods to perform descriptive analysis on the bikeshare data from three major U.S. cities - Chicago, Washington, and New York City - to display information such as most popular days or most common stations.

### Running the program

You can input 'python bikeshare.py' on your terminal to run this program. I use Anaconda's command prompt on a Windows 10 machine.

### Program Details

The program takes user input for the city (e.g. Chicago), month for which the user wants to view data (e.g. January; also includes an 'all' option), and day for which the user wants to view data (e.g. Monday; also includes an 'all' option).

Upon receiving the user input, it goes ahead and asks the user if they want to view the raw data (5 rows of data initially) or not. Following the input received, the program prints the following details:

*   Most popular month

*   Most popular day

*   Most popular hour

*   Most popular start station

*   Most popular end station

*   Most popular combination of start and end stations

*   Total trip duration

*   Average trip duration

*   Types of users by number

*   Types of users by gender (if available)

*   The oldest user (if available)

*   The youngest user (if available)

*   The most common birth year amongst users (if available)

Finally, the user is prompted with the choice of restarting the program or not.

https://github.com/fagrahmed12/bike-Share/tree/master


Some of My Projects