Sharad 's Projects
This is the Major Project carried out during my final Semester as part of my B.Tech. This project uses the cropped image set of UTKFace dataset for age and gender detection. The technique used is Convolutional Neural Networks (CNN) and the basic architecture is inspired by VGG-16 model.
Batch ETL pipeline using Apache Sqoop, Apache PySpark, Amazon S3 and Amazon RedShift to analyze ATM withdrawl behaviours to optimally manage the refill frequency.
Covid ETL Project using Azure Data Engineering Stack
Multiple Regression model building with Sklearn and statsmodels and analysis of relevant predictors using P-values and VIF
Credit card fraud detection of real time transaction data
Extensive EDA Case study of Loan applications of customers based on various factors and identifying the trends in Defaulters and Non Defaulters
NYC Taxi data analysis using Mapreduce
A collection of the DSA problems solved in Java as part of DSA course by Coding Ninjas
Scoring Leads for an Ed-Tech company to enable higher leads conversion
This notebook uses different classification models to predict how many customers of a bank will pay the loan and how many will be defaulters.
This is the mini project carried out during the summer of 2020 as part of the requirement for B.Tech curriculum. A convolutional autoencoder model for denoising images . Here I have used the Mini-Mias mammography dataset
Case Study on a Movie Production House using SQL
Demand Forecasting is the process in which historical sales data is used to develop an estimate of an expected forecast of customer demand. I worked on the Store Item Demand Forecasting dataset available at Kaggle (https://www.kaggle.com/c/demand-forecasting-kernels-only) . The dataset consists of 10 stores and 50 items and their respective sales . In my project i used the plotly and seaborn visualization libraries for plotting which are an excellent tool to get insights into the data. Feature engineering was performed to get the right features for predicting the sales.I used the following ML models : Gradient Boosting Regressor ,Decision Tree Regressor ,Linear SVR ,Random forest Regressor and compared the performance . Finally, deep learning implementation is also done using LSTM.
Titanic dataset consists of the passenger details on the Titanic and if they survived or not. Different classifiers are used to predict the survival status of the passengers in the train set and their accuracy noted and the model with best classification accuracy is used to predict the survival status on the test set.
Repository of algorithms and data for public transit routing
This project depicts the visualization of Uber pickup data in New York city and uses the Neighborhoods JSON file of New York city and ML algorithms to predict the no. of pickups in each neighborhood.