Data Science Portfolio
This section contains portfolio of data science projects completed by me for academic, self learning, and hobby purposes.
For a more visually pleasant experience for browsing the portfolio, check out jameskle.com/data-portfolio
-
Recommendation Systems
-
Transfer Rec: My ongoing research work to incorporate transfer learning into the design of recommendation systems.
-
Movie Recommendation: Designed 4 different models that recommend items on the MovieLens dataset.
Tools: PyTorch, TensorBoard, Keras, Pandas, NumPy, SciPy, Matplotlib, Seaborn, Scikit-Learn, Surprise, Wordcloud
-
-
Machine Learning
-
Trip Optimizer: Used XGBoost and evolutionary algorithms to optimize the travel time for taxi vehicles in New York City.
-
Instacart Market Basket Analysis: Tackled the Instacart Market Basket Analysis challenge to predict which products will be in a user's next order.
Tools: Pandas, NumPy, Matplotlib, XGBoost, Geopy, Scikit-Learn
-
-
Computer Vision
-
Fashion Recommendation: Built a ResNet-based model that classifies and recommends fashion images in the DeepFashion database based on semantic similarity.
-
Fashion Classification: Developed 4 different Convolutional Neural Networks that classify images in the Fashion MNIST dataset.
-
Dog Breed Classification: Designed a Convolutional Neural Network that identifies dog breed.
-
Road Segmentation: Implemented a Fully-Convolutional Network for semantic segmentation task in the Kitty Road Dataset.
Tools: TensorFlow, Keras, Pandas, NumPy, Matplotlib, Scikit-Learn, TensorBoard
-
-
Data Analysis and Visualization
-
World Cup 2018 Team Analysis: Analysis and visualization of the FIFA 18 dataset to predict the best possible international squad lineups for 10 teams at the 2018 World Cup in Russia.
-
Spotify Artists Analysis: Analysis and visualization of musical styles from 50 different artists with a wide range of genres on Spotify.
Tools: Pandas, NumPy, Matplotlib, Rspotify, httr, dplyr, tidyr, radarchart, ggplot2
-
Data Journalism Portfolio
This section contains portfolio of data journalism articles completed by me for freelance clients and self-learning purposes.
For a more visually pleasant experience for browsing the portfolio, check out jameskle.com/data-journalism
-
Statistics
-
Machine Learning
-
Deep Learning
-
The 8 Neural Network Architectures ML Researchers Need to Learn
-
The 5 Deep Learning Frameworks Every Serious Machine Learner Should Be Familiar With
-
The 5 Computer Vision Techniques That Will Change How You See The World
-
Convolutional Neural Networks: The Biologically-Inspired Model
-
Recurrent Neural Networks: The Powerhouse of Language Modeling
-
The 7 NLP Techniques That Will Change How You Communicate in the Future
-
The 3 Deep Learning Frameworks For End-to-End Speech Recognition That Power Your Devices
-
The 5 Algorithms for Efficient Deep Learning Inference on Small Devices
-
The 4 Research Techniques to Train Deep Neural Network Models More Efficiently
-
The 2 Hardware Architectures for Efficient Training and Inference of Deep Nets