Regression & Predictive Modeling and Analysis of Bitcoin Hashrate
Forecasting & TIme Series Analysis
The primary objective of this study was to develop a dependable and precise prediction model to forecast alterations in Bitcoin's hash rate.
Marathon Digital Holdings' aim was to attain a comprehensive comprehension of hash rate fluctuations, signifying the collective computational capacity employed for mining and transaction processing on the Bitcoin network.
Achieving this objective will empower Marathon Digital Holdings to adopt data-guided decisions, enhance mining operations, and elevate their overall effectiveness in the realm of Bitcoin trading.
​
Skills: Tableau Desktop · Matplotlib · Microsoft Excel · Python (Programming Language) · Machine Learning · Data Analysis · Statistical Data Analysis · Data Visualization · Jupyter
​
Predictive Analytics
Machine Learning, Jupyter Notebooks
The project contains the implementation of multiple small-scale projects related to Predictive Analytics where each step of data analytics is executed, right from Exploratory Data Analysis, Data Cleaning and Visualizations, Model Building, and Recommendations based on findings.
Multiple machine learning algorithms and models are implemented and compared to analyze the best-fit model based on the business context.
The implementation of each of the projects is developed using Python Programming Language using Jupyter Notebook and Google Colab.
​
Skills: Microsoft Excel · Machine Learning · Data Analysis · Statistical Data Analysis · Data Visualization · Jupyter · Data Science
​
​AWS Cloud Architecture
Amazon Web Services
The project dealt with the implementation of an architectural diagram for a medical company that aims to build its application on the AWS cloud platform. The final architecture build was based on the customer requirements along with the various AWS services. The VPC was divided into two parts, i.e., the development and production which are in two availability zones. The various Amazon services included are Amazon CloudTrail, Amazon Cognito, Route 53, Amazon CloudFront, CloudWatch, and IAM.
​
Skills: Amazon Web Services (AWS) · AWS Cloud
​
Applications of Artificial Intelligence
Natural Language Processing, AI
The project contains the implementation of various applications of Artificial Intelligence such as Distributed AI applications, Natural Language Processing applications, Vision AI applications, and Robotic AI applications. The implementation of each of these applications was developed using Python Programming Language using Jupyter Notebook and Google Collab.
​
Skills: Natural Language Processing (NLP) · Big Data Analytics · Artificial Intelligence (AI)
​
Motor Vehicle Collision Crash Analysis
Data Analysis, Machine Learning
The data for the vehicle collisions and accidents have been collected which helps in the analysis of this data such that the incidents can be prevented, and safety measures can be taken accordingly. The goal here was to identify the factors that are causing vehicle collisions and recommend which factors would lead to a severe incident or a non-severe incident such that appropriate measures can be taken to order to prevent crashes and accidents and ensure the safety of the public. The variable that helped with the prediction or classification of the vehicle collisions was the contributing factor of the vehicle and the number of persons killed or injured.
The various data analytical steps performed were exploratory data analysis, data cleaning, data visualization, dashboard analysis, pre-modeling steps, model building, and providing recommendations and findings based on the analysis.
​
Skills: Python, Jupyter Notebook, Data Visualization (Tableau), Machine Learning (Regressor Models)
​
Student Risk Analysis (Student 360)
Machine Learning, Risk Analysis, Deployment
Student 360, a project of the sponsor company, SwatCloud & Siliconvalley4u, dealt with analyzing student performance based on various external factors to determine the student dropout rate and predict the CGPA of the students.
The dataset obtained from the sponsor was analyzed and visualized to get a better understanding of the data and field values. The data analysis workflow consisted of Data Collection and preparation, Data Exploration (Exploratory Data Analysis), Data Cleaning, Data Visualization, Model Building (Predictive Analysis), and Deployment.
The developed system is able to classify students into the eligible or not eligible category based on their overall performance. It can also predict the CGPA of the BTech course for the new students to understand whether the students will drop out of the course or not.
​
The project code can be found here:
https://github.com/rrambhia22/Student360
​
Projects
Analyzing the number of deaths by risk factors in R
R Programming, Statistical Analysis
Exploratory Data Analysis, Hypothesis Testing, and Regression Model were performed on the dataset of the number of deaths by risk factors to analyze and examine the relationship between the variables of the dataset. The analysis done on this dataset would help in determining which risk factor would highly affect the death rate and what is the death rate due to these factors in each country.
​
Skills: Microsoft Excel · R (Programming Language) · Data Analysis
​
Automation Testing
Java, Selenium
A Maven project to book the flight on yatra.com was created with a Hybrid Framework based on Page Factory. This was an automation testing project using Selenium WebDriver along with various other libraries. It included testing the site having different functionalities for flight booking.
​
Skills: Java, Selenium, Hybrid Framework
​
Software Tool to measure the effect of teacher presence in discussion forum of MOOCs on Peer Learning
Machine Learning
Massive Online Open Courses are online courses that provide interactive user forums or discussion forums that build a community for students and professors where it becomes important to measure the effect of teachers' presence in these forums.
The learning in the forum of the IITB course has various categories and due to the huge number of participants involved in the course, it makes manual classification impractical.
A classification model was built which helps in automating the process of categorizing the input messages of the discussion forum into their type of learning. After the classification, analysis was done to check whether teacher intervention resulted in a higher-order level of learning.
The proposed system helped in analyzing how the presence of teachers in MOOCs contributes to the quality of students' learning since it is difficult to quantify this manually.
​
Skills: Leadership · Teamwork · Communication · Microsoft Excel · Python (Programming Language) · Machine Learning​
Automated News Classification
Machine Learning
As there are a lot of sources that generate an immense amount of daily news and the demand for information by users has been growing, there is a need to classify the news reports & articles into their respective categories.
A model was therefore built for automating the process of news classification in order to classify the news headlines into their respective categories where the Neural Network model (DNN) was implemented along with Natural Language Processing.
The model was built for both English and Regional language (Gujarati). The input to the model can be either text input or audio input and when given an audio input, it would convert into textual format i.e. speech to text classification, and then classify.
There were various stages performed to complete this task such as Data Preparation, Model Building and Visual Analysis.
​
Skills: Microsoft Excel · Python (Programming Language) · Machine Learning
​
IITB MOOC
Python & Data Analysis
Massive Online Open Courses are online courses that are aimed at large-scale interactive participation and open access via the web.
The objective was to gauge the ability to work with real MOOC data and learning analytics. Analysis of the MOOC forum course of IIT Bombay was performed using methods like Data Cleaning, Data Extraction, and Visualization. The methods used for the extraction of data were manually coded which were then used for the visual analysis.
This was the preliminary analysis before developing a 'Software Tool to measure the effect of teacher presence in the discussion forums of MOOCs on Peer Learning'.
​
Skills: Teamwork · Communication · Microsoft Excel · Python (Programming Language)
​
Online Book Buying System
Web Development
A small-scale website for Online Book Buying was implemented. The website included the process of ordering books from a book bank which was automated.
It consists of various events such as the Home Page, Registration Page, Login Page, and Book Page. PHP and Xampp were used for the backend framework which stored the user details.
​
Skills: SQL · HTML · Cascading Style Sheets (CSS) · PHP · Web Development
​