Data is an exact thumbprint of the world's behaviour and psychology at any given moment.
With data, I get to know what the world is thinking and feeling and what it will do next.
I used some fancy maths and Google trends to uncover relationships between people's online searches and their economy. I found out that U.S. states with higher rates of infant mortality Google for bad credit, loans, and information on sexually transmitted diseases.
I wrote a paper about it.
I wanted to know if we can quantify aspects of academic writing that correlate with how a paper is received. I found out that scientific papers with shorter titles recieve more citations.
I wrote a paper about it. It made the front page of Reddit for a day. It was featured in:
Analysing huge amounts of Google data is hard when you have to download it by hand.
I made a Python library to download Google data that’s protected by a login.
Airtasker helps you to realise the full value of your skills.
I’m responsible for building the machine learning pipeline and algorithms at Airtasker. I've built systems to:
- Automate platform moderation. Many things posted to the platform do not fit with our community guidelines. We have a set of algorithms as our front-line defense removing undesirable content.
- Categorise tasks. Hundreds of thousands of tasks are posted to Airtasker. We automatically classify the to help our users find what they are looking for.
- Recommendation systems. Due to the shear volume of tasks on the platform, we use recommendation algorithms to help surface the right tasks to our users.
- Pricing algorithms. Understanding how much a task is worth is important for both sides of the transaction.
- Productionising machine learning algorithms. Developing an algorithm is only a small fraction of delivering a machine learning solution. Most of the battle is productionising the models in such a way that we can rely on them, monitor them, and push out updates quickly.
When you’re learning maths, Mathspace is the right help at the right time.
I built tools to help Mathspace understand their data and create more value for their customers and the company. I developed the next generation of adaptive learning algorithms to further automate the learning experience. Kids are now learning independently.
I used online data such as Google search logs to predict real world behaviour like future stock prices. Highlights include:
- Engineered various web tools in for business stakeholders to analyse large datasets.-Lead the development or had a hand in developing any software that was used outside of our team.
- Created a multitude of data visualisations and infographics to showcase my own work and the work of the team.
I developed a rigorous mathematical framework for traditional methods of trading. I used statistical analysis to dissect and understand how the methods generated profits.