Dan Bouchard's Projects
A project to predict the price of an Airbnb listing from tabular, text and image data
In this lab, you'll implement an industry grade data collection pipeline that runs scalably in the cloud.
A project to develop and train a Facebook Marketplace Search Ranking system which uses is a trained multimodal model accepting both text and image data in order to generate vector embeddings in order to make recommendations for a user searching for a product to buy.
Football Match Prediction Classifier using data from over 30 years from the top European leagues to predict future match results. The project involves: EDA, data cleaning, feature engineering, feature selection, training and optimising multiple classification models for the best accuracy.
Leave One Feature Out Importance
A data processing pipeline for handling Pinterest data. Built a batch and a streaming pipeline using Kafka and Spark. The whole pipeline can process and store large amounts of data and compute accurate metrics using both historical and recent data.
SQL Exploratory Data Analysis and Visualisation of the Austin Bike Share Trips dataset available in Google BigQuery
Data visualisation using Tableau of Covid Airport Traffic data over the year 2020.