Welcome to the landing page for Tom’s personal github account.
Tom Croshaw is a catastrophe research analyst turned data scientist with a keen interest in the natural environment, science and of course… Data!
Here you can review example projects with the code and analytics used to produce the insights. Please reach out to me on LinkedIn if you’d like to know more…
Can Social Data Predict Severe Natural Catastrophes?
April - June 2017
This project is an investigation into how Twitter data can be used to predict major natural catastrophes. Specifically, can social data help us determine whether:
- There will be a large financial loss
- A significant number of insurance claims will be received
- A large area will be impacted by the event
This investigation was my Data Science Immersive Capstone Project at General Assembly, Sydney. Please follow the link below to view the progress of the analysis:
Investigation Blog
Property Sales Price Regression Analysis
September 2017
This project is an investigation into whether property attribute data can be used to predict the sale price of a home. It approaches the analysis from the perspective of an property investment company looking to optimise returns.
The analysis uses the Ames, Iowa Housing dataset and compares the impact of fixed, permanent property features on the sales price to those that can be changed i.e.: renovation potential.
Property Sales Analysis Report
Predicting Salaries of Data and Research Jobs
February 2018
This analysis investigates which skills, industries and positions of data and research jobs pay the highest salaries.
The analysis uses data scraped from a job seeking aggreator, natural language processing and machine learning to determine which features predict the highest salaries, and which are most needed for data science roles. The python notebooks for web scraping and the analysis are provided to guide you through the analysis and conclusions.