Vishal Barad
Data scientist || Data engineer || Pyspark || Python || SQL || 1x Airflow certified || Databricks || Azure
Need a Data scientist, Data engineer and Python Developer to build the solution?
I'll help you create a well-structured, scalable solution with my python, machine learning and deep learning, pyspark skills as scalable web app with django and django rest framework skills that’s perfect for you.
I'm a Data scientist and Data engineer with experience in various technologies which includes Machine learning, Deep learning, NLP, cloud technologies, Software engineering etc. Skilled engineer with around 2 years of experience and demonstrated history of good architecture and technical knowledge in Data science and engineering.
My toolkit includes python, pyspark, machine learning, NLP, deep learning, django/django rest, flask, PostgreSQL, Airflow, azure, databricks as well as tools like Heroku for deployment. I'm also always happy to learn new tools!
IPL match score prediction (Freelance)
The purpose of this project is to predict the first inning IPL match score based on Venue, Batting team, Bowling team, Runs, Overs and Wickets etc.
• Gathered dataset. Dataset used by this project is ‘ipl.csv’ downloaded from kaggle. After downloading and importing dataset(in jupyter Notebook) i did data cleaning first like dropping some unnecessary columns, handling missing values, performing one-hot encoding on categorical variables.
• Divided dataset into dependent and independent features. (Independent features=Venue,runs,wickets,etc.. | Dependent feature=total)
• Dropped first 5 overs data in every match because first 5 overs are powerplay over ,soi just ignored it.
• splitted data into training data and testing data.And then i perform model selection in which i chose ‘Multiplelinear’, ‘Ridge’, ‘Lasso’, ‘Decision tree’ and ‘Random forest’ regression algorithm and count accuracy score.
Python, Machine learning regression algorithm, Numpy, Pandas, Joblib, Flask, Heroku, HTML, CSS
Movie recommendation (Self Project)
The purpose of this project is to recommend top 10 similar movies which user watched before. This is project based on ‘model based collborativefilering’ ml algorithm.
• Gathered dataset. Dataset used by this project is ‘movies.csv’ and ‘ratings.csv’ downloaded from movielens dataset ‘https://grouplens.org/datasets/movielens/latest/’. After downloading and importing dataset(in jupyter Notebook) i did data cleaning first like dropping some unnecessary columns, handling missing values, Merging two datasets ‘movie’ and ‘rating’ on ‘movieId’.
• created ‘pivot table’ and then ‘sparse matrix’.
• fitted ‘sparse matrix’ using ‘NearestNeighbors’ model.
• used ‘kneighbors()’(which returns distance and indices) method of ‘nearestneighbor’ model to predict the similar user. And just predicted top 10 similar movies.
• Saved model using ‘joblib’ library. After that I creted UI in flask
• Deployed on ‘Heroku’.
Python, Machine learning regression algorithm, Numpy, Pandas, Joblib, Flask, Heroku, HTML, CSS, Pycharm
Smiley face detection (Self Project)
This project is made using deep learning CNN neural net and opencv. This project detects whether we are happy or neutral.
• Gathered dataset. Dataset used by this project is downloaded from kaggle.
• Data augmentation to generate more data to avoid overfitting
• Applied CNN’s different-different layers like conv2d, maxpolling, dropout, filter at last flatten layer.
• Compiled model based on optimiser and accuracy matrix
• Saved model using ‘joblib’ library. And used opencv to detect face and predict output.
Python, Deep learning, Numpy, CNN, Opencv, Pandas, Joblib, Flask, Heroku, HTML, CSS, Pycharm
Face mask detection (Self Project)
This project is made using deep learning CNN neural net and opencv. The purpose of this project is to detect whether person has wear mask or not.
• This is project based on CNN model. Dataset used by this project is 'https://drive.google.com/drive/folders/1P3gIgFUMbdl5tSqx1pK385bz4t0mxEvW?usp=sharing'. Dataset conatins train, test and validation data in each there are two class 'with_mask' and 'without_mask'. After downloading and importing dataset(in Google colab) i did data image augmentation first like rescale image, rotate image, flip image, zoom image etc.
• After that i just compile model using 'Adam' optimizer with learning rate=0.01, loss='sparse_categorical_crossentropy'.
• After that i just fit data using 50 epochs.At the end i got 98% validation accuracy and 96% training accuracy.
• Then again i just saved model as 'final_Face_mask_scratch.h5'. After downloading model i made 'prediction.ipynb' file using Opencv to predict output.
Python, Deep learning, Numpy, CNN, Opencv, Pandas, Joblib, Flask, Heroku, HTML, CSS, Pycharm
Rasa product chatbot (Freelance)
This project is made using NLP rasa chatbot. The purpose of this project is to assist best product based on user's question and answer.
• Developed simple RASA chatbot
• Designed workflow
• Extract keyword from user answer using NLP
• Recommended best product to user
Python, Pycharm, mysql, machine learning, NLP, RASA chatbot
Xplotue (Infozium solution PVT LTD)
Xplotur is an online travel platform that tries to exceed traveler’s expectations with absolute determination and commitment. It offers travel planning, Itinerary design, transportation facilities, hotel booking, Flight booking services.
• Developed Hotel module
• Communicated with hotel API provider
• Integrated hotel API with frontend
php, VS code, mysql, wordpress
Voiceme project (Kintu designs PVT LTD)
VoiceMe is an online bank application in which user can do bank truncation with their voice by just communicating with rasa chatbot (same as Siri).
• Developed simple RASA chatbot
• Designed workflow
• Perform text to speech using GTTS(Google Text-to-Speech) library
• Extract keyword from user voice using NLP
• Perform speech to text using SpeechRcognition library
Python, Pycharm, mysql, machine learning, NLP, RASA chatbot
Simplyloose (Kintu designs PVT LTD)
Simplyloose is the IOT based online personal training platform for Fitness experts- Doctor, Dietitians Gym, Yoga Trainers, Physiotherapists to better connect to their customers when training them online or in-person.
• Scraped data using BS4 and Scarpy
• Designed workflow
• Perform data pre-processing on scraped data
• Used scraped data to built machine learning model
• It was a classification problem so I have used random forest classifier algo to predict diet and recipes.
• Applied natural language processing to extract keyword of user input during chat
Python, Pycharm, mysql, machine learning, web scarping, Django
Tamil Nadu e-Governance Agency – SFDB, Govt. of Tamil Nadu (Vedity software pvt ltd and KPMG)
One of the core functions of the this project is to deliver the public services to its citizens. With increased resident expectations, the government is expected to deliver the public services in a more convenient manner. To meet these expectations, the government intends to provide timely and convenient delivery of public services, social benefits and improve the efficiency of public service delivery.
• Data preprocessing pipeline design
• Wrote pyspark script for data ETL. We used pyspark here because the size of the data is around 7 crore.
• Defined pipeline in using airflow
Python, pyspark, airflow, pandas, machine learning algorithm, postgres sql
Blogs
Data Preprocessing Using Pyspark (Part:1)
Apache Spark is a framework that allows for quick data processing on large amounts of data. . . . . .
Unsupervised learning algorithm
In unsupervised learning algorithm we don't supervise the model, but we let the model work on its own to discover information that may not br visible to the human eyes. . . . . . .
Skills & Tools
Python
language
Machine
learning
Deep
learning
Pyspark
tool
Microsoft
azure
Apache
airflow
SQL
Postgresql, Mysql
Azure
databricks
Django
| Django rest framework
Web
scraping
Natural language
processing
Pandas
tool
Data scientist | Data engineer
Vedity software pvt ltd
Ahmedabad
Sept 2021 - Current
Data scientist
Kintu design pvt ltd
Surat
May 2021 - Sept 2021
Full stack developer
Infozium solution pvt ltd
Surat
Feb 2021 - May 2021
Data science and python developer Freelancer
freelacer.com and upwork.com
Surat
Nov 2020 - Sept 2021
Certifications
Testimonials
Nshini
freelance - July 2021
Build me a chatbot using Rasa or any open source programs
The work was professionally done within the requirements and budget. It was easy to communicate with him and he answered all my queries promptly. I will definitely recommend him.
Skills: Python Programming, Software Architecture, Chatbot
Soham A.
freelance - Nov 2020
Boosting regression algorithm solution on paper
Good Job done Vishal. Project completed successfully and happy to work with him in future.
Skills: Python, Algorithm, Report Writing, Machine Learning (ML), Statistical Analysis
Ashish K.
freelance - Dec 2020
Prediction of English Premier league using Artificial Intelligence
Good job Vishal , keep it up !
Skills: Machine Learning (ML), React.js, Artificial Intelligence, Blockchain, Ethereum
freelance - Nov 2020
Machine learning classification -- 3
have good knowledge of machine learning and has experience in coding too. Delivered the project on time.
Skills: Python, Matlab and Mathematica, Machine Learning (ML), Deep Learning
Fahmi E.
freelance - Oct 2020
Expert Django Rest Framework Required for Clean Code Design Review and Refactor - Hourly Consultation
An excellent django developer. Vishal helped me restructure my django code infrastructure. Fast and responsive. Straight to the point.
Skills: Python, Django, RESTful API
Get in touch
If you like my profile and according to you I'm a best employee who fullfill your work then get in touch with me or you can directly mail me on [email protected].
