Kewal Tadas

Kewal Tadas

Master's in Data Science at Seattle University

Sitemap

A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.

Pages

About me

Posts

Blog Post number 4

less than 1 minute read

Published:

This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.

Blog Post number 3

less than 1 minute read

Published:

This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.

Blog Post number 2

less than 1 minute read

Published:

This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.

Blog Post number 1

less than 1 minute read

Published:

This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.

portfolio

A Decision Tree Analysis on Drug Use and Health

Published:

This report contains the analysis of factors that are associated with youth drug use, using decision tree models on survey data from the National Survey on Drug Use and Health. The dataset contains detailed information on various aspects of respondents’ lives, including demographics, youth experiences, and substantial drug use. The study explores three problem types: binary classification for marijuana use, multi-class classification for frequency of marijuana in past year, and regression for first ever use of marijuana. Decision trees and some ensemble methods are used to build predictive models for each problem type. The results suggest the impact of different demographic variables and youth experience variables on various substantial youth drug uses. The findings of this study have practical implications for public health interventions aimed at reducing youth drug use.

Bird Sound Classification using Neural Networks: Transfer Learning and Challenges

Published:

In this code, we perform bird sound classification using a neural network to predict bird species based on their sounds. We used data from the Birdcall competition, selecting 12 bird species. We are trying to build both a binary classification model for two of the species and a multi-class classification model for all 12 species. A crucial aspect of our project involved exploring transfer learning. To execute this, we implemented a pre-established neural network model that underwent modifications tailored for our particular data. Our investigation revealed that the neural network exhibited an acceptable level of performance. Nevertheless, we encountered some limitations & obstacles that need to be overcome in further research.

Comparative Analysis of Fuel Efficiency Trends by Origin (1970-1981)

Published:

The visual representation employs line charts differentiated by colors to display the average miles per gallon over various years, categorized by origin (Europe, Japan, USA). This encoding choice is effective because line charts are intuitive for illustrating time series data and tracking changes over time, allowing for easy comparison between the different origins.

An Interactive Visualization of Chicago’s Cultural Landscape

Published:

Mapping Chicago’s Cultural Landscape offers an interactive guide to the Windy City’s museums, zoos, and aquariums. This project harnesses the power of Folium mapping to visually chart the locations of cultural icons, each marked with distinct, color-coded pins. It serves as an essential resource for both residents and visitors to discover and connect with Chicago’s rich educational and cultural offerings. Through this intuitive map, the project aims to enhance the accessibility and exploration of the city’s diverse institutions.

Classification of Housing Ownership Using SVM

Published:

This report applies Support Vector Machines (SVM) to classify housing ownership (owned or rented) based on demographic and economic variables. The analysis uses US Census data accessed through IPUMS USA, with pre-processed data grouped by households and irrelevant variables removed. SVM models with three kernels (linear, radial, and polynomial) are employed to capture the relationship between variables and occupancy status. Accuracy is assessed with varying regularization parameter (C) values, and decision boundary plots visualize the classification results. The findings showcase SVM’s predictive ability, as all three models achieve high accuracy in predicting occupancy status.

Connect-4 Outcome Analysis using supervised and Unsupervised Algorithms

Published:

This project aims to investigate various supervised and unsupervised algorithms for analyzing the Connect-4 game dataset. The dataset consists of legal game positions in Connect-4 where neither player has achieved victory yet, and the next move is not predetermined. The project commences with data visualization, including a bar plot representing the distribution of outcomes and a heatmap illustrating the game board positions. To pre-process the data, categorical features are encoded using one-hot encoding. Several supervised algorithms such as Logistic Regression, Decision Trees, Support Vector Machines, and Neural Networks are employed, alongside unsupervised learning techniques like K-means clustering. The performance of the supervised learning algorithms is evaluated based on accuracy, while the unsupervised clustering approach is evaluated using metrics such as the silhouette score and completeness score. This project aims to provide valuable insights into the efficacy of different techniques for analyzing the Connect-4 dataset.

Analysis of Crime Trend in Seattle

Published:

The analysis presents a comprehensive analysis of crime distribution and trends in the city of Seattle for the years 2008 to 2022, emphasizing year 2022 to understand the present crime dynamics in Seattle, by utilizing a robust dataset obtained from the Seattle Police Department. The dataset is a rich amalgamation of information, detailing the types of offenses, the locations where they occurred, the times at which they took place, and other critical variables that give a multi-dimensional view of criminal activity within the city.

GYM Fitness Management System.

Published:

Boutique gyms and health club startups often face the challenge of managing their operations effectively while keeping costs low. Our application offers a graphical user interface (GUI) designed specifically for these businesses, prioritizing simplicity and affordability.

Stroke Prediction

Published:

A stroke is a cerebrovascular disorder that causes damage to the normal supply of blood to the brain, according to the World Health Organization (WHO). Strokes account for approximately 11% of all deaths worldwide. The condition may lead to death or long-term disability if not treated properly. The cause of strokes can be hemorrhagic or ischemic. Both hemorrhagic and ischemic strokes can occur together. Ischemic strokes are those caused by clots in the blood vessels, while hemorrhagic strokes are those caused by ruptured blood vessels. . In general, 85 percent of strokes are classified as ischemic.

publications

talks

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.