Data Visualisation & Creative Coding



For data analysis: I mainly use Python (NumPy, SciPy, pandas, matplotlib, scikit-learn); IDL, SQL, and UNIX languages (e.g. Bash, AWK) and shell scripts (e.g C–shell for wrappers and batch jobs). For code management & issues tracking: I use git version control with GitHubBitBucket, SourceTree (Git GUI); Jupyter Notebooks, Jupyter Lab, and Spyder IDE. Specific data visualisation libraries & tools: include, matplotlib, ggplotPlotly, BokehD3.js and dimple.js javascript libraries, Carto, TimelineJS, and Odessy. For machine learning: I use scikit-learn (e.g. for classification, clustering and regression), and NLTK (e.g. for language processing). For web development: I use HTML5, CSS, and JavaScript libraries (JQuery), JSONGitHub Pages (for hosting);  templated.co and HTML5UP! (for web templates); Atom and LightTable (for code editing), GIMP (GNU Image Manipulation), and Dreamweaver

A few years ago I started creating small projects outside of my day-to-day work, and began writing them up as blog posts. The intent was to demonstrate what can can be created with just a handful of coding tools in your kit. Creating projects (no matter how small) is an excellent way to learn a new programming language, implement a new library or package, and to get your hands dirty with tools you are not familiar with.  It's also a great way explore various data analysis techniques, and data visualisation tools. I’m also big fan on using ready-to-use web tools. These can be really useful for outreach projects, or when I want to get something visual up and running fairly quickly and don’t necessarily want (or need) to build it from scratch.

[Selected projects]

Data Science & Interactive Visualisations

Exploratory Data Analysis  »  Creative Machine Learning  »  D3.js experiments

shoppar_attention_whiskers.jpg

Understanding Customer Behaviour: Predicting Attention Span

Visualising machine learning models and delivering business insights. Working with Shoppar as a Pivigo Data Science Fellow

Creating colour palettes from images

A Jupyter Notebook tutorial that uses Python's scikit-learn & matplotlib package

Machine Learning Tutorial using SDSS quasar data

A Jupyter Notebook tutorial that explores Machine Learning methods (clustering, regression) using Python's scikit-learn & matplotlib package

Astrophysics to Data Science

Visualising the career paths of astronomers turned data scientists – D3.js sequences starburst.

Exploring research data interactively

Creating an interactive plots of the Atlas3D dataset using the dimple.js javascript library

Visualising the UNIX/Python Software Carpentry workshop

The D3.js sequences sunburst was used to
explore the diversity of workshop participants

Research Backgrounds of Insight Data Science Fellows

Exploring the research profiles of Insight Data Science Fellows using the d3js co-occurrence matrix & trulia heat map

Visualising Research Data

Astronomy data presented in scientific journal papers, conference talks, and research meetings.

 
 

[SELECTED PROJECTS]

Data Storytelling & Visual Communication

Exploratory Data Analysis  »  Interactive Visualisation  »  Science Communication

Visualising an art exhibition

A may contain data project that includes both tactile and interactive web (coming soon) visualisations

A Short History of Gravitational Waves

An interactive TimelineJS visualisation created for OzGrav – the ARC CoE for Gravitational Wave Discovery

Interactive UTMOST Commissioning Timeline

An interactive TimelineJS visualisation created for the UTMOST Telescope

Find your next AstroTech conference

A visualisation for techsavvyastro.io built using the Carto mapping platform. Also uses Google Spreadsheets, SQL & custom CSS

Mapping the rise of Data Science Institutes around the world

A visualisation built using the Carto mapping platform. Also uses Google Spreadsheet & SQL, and custom CSS

Interactive Map of Optical & Radio Telescopes

A visualisation built using the Carto mapping platform. Also uses Google Spreadsheets, SQL & custom CSS

 

Web Experiments

Science Communication  »  Citizen Science Projects  »  #dotastro Experiments

A website that makes decisions so I don't have to

Built using GitHub, templated.co, HTML5, CSS & JQuery

techsavvyastro.io

A Squarespace website that empowering researchers to redesign their careers. Logo design using iconfinder & GIMP

#dotastro Hack Graveyard

Need a hack idea? revive one of ours! Forgotten or failed hacks from .Astronomy. Built using GitHub, templated.co, HTML5, CSS & JQuery

.Astronomy Hacks – Image Gallery

A visual feast of hack day projects from the past eight years of .Astronomy conferences. Built using GitHub, templated.co, HTML5 & CSS

project_apollo_stories.jpg

Project Apollo – Data Driven Stories

A demo citizen science project to ecreate the stories from the Apollo missions. Built using the Flickr API, Python, TimelineJS, and the Zooniverse Project Builder