I am a senior data science consultant working in strategy consulting at Port Jackson Partners in Sydney, Australia.
I undertook my PhD in computational biology at the Australian National University where I developed image analysis software to track honeybee behaviour using machine learning.
I am accustomed to working independently to solve commercial problems in a short timeframe
I am a data science consultant at Port Jackson Partners (PJP), a management consulting firm that work with top-tier Australian and multi-national organisations to solve complex commercial problems. As the only data scientist in the traditionally strategy-focused firm, I am responsible for the design and delivery of the analytics and technical component of projects in order to generate insights that can be used to solve significant business challenges.
My role at the firm is to liaise with partners and teams in order to independently solve any data and analytics challenges that they face on typically high-stakes, fast-paced projects. This problem solving process will typically involve me identifying, acquiring, cleaning, linking, and analysing multiple datasets. I will frequently utilise web scraping techniques in order to augment the datasets I have access to on a project.
I create value that has a lasting impact on the teams and clients that I work with
Through delivering consulting projects at PJP, I have acquired considerable domain expertise and developed curated datasets in order to solve data challenges that are of particular value to the client and the firm. For instance, I maintain a curated version of the data from the National Energy Market (NEM). This has been a valuable, enduring asset for the firm that has been used on multiple projects for both Industry and Government clients. I have used this dataset to solve a range of problems, ranging from calculating the potential revenue of new renewable infrastructure, through to running simulations based on historic pricing data to calculate the revenue that new battery assets in the NEM would generate.
During my time at PJP, I have proactively acquired skills which have proven to be an asset on multiple projects. For instance, I currently have considerable experience working with GIS data utilising Python’s GeoPandas library. This has been an asset on multiple projects as there is frequently a spatial dimension to most data I acquire (or I can geocode address data using the Google Maps API). Furthermore, the spatial datasets released by the ABS as part of the 2016 census makes it extremely easy to group coordinates (such as classifying which suburb or local government area they fall into) or to link the data up with demographic data from the census.
Another skill I have acquired as part of my work at the firm has been that of building interactive visualisation tools and dashboards utilising Python’s Bokeh library. I started developing these tools because I noticed team members who could not program needed to work with the large datasets that I had been manipulating. Aside from their value within the team, these interactive tools have also played a valuable role in client problem-solving workshops and been a key deliverable to the client at the completion of the project. Some examples of tools I have built include visualisations to explore granular customer profitability for energy retailers, the future profitability of thousands of non-government schools based on government legislation and multiple scenarios, and the relationship between schools and the before and after school services available nearby.
I am passionate about using my programming and data skills to solve analytical problems
I really enjoy the challenge of being staffed on a new project and having to rapidly identify which datasets I can clean and link to achieve the outcome that the client and team require. I have become extremely effective at accomplishing these tasks across multiple projects over the past two years.
While I am used to the rapid pace of consulting, I am also comfortable working on longer term projects that require deep technical expertise. During my PhD, I wrote image processing software and trained machine learning models to distinguish between patterns in tags I designed. As part of my research, I utilised this software to track honeybee behaviour in the hive and quantify their behaviour. I have also conducted a range of side projects in my spare time using the Raspberry Pi computer to process video in real-time in order to quantify the activity levels of honeybee and native bee hives out in the field.