Visualising Category Recoding Graphs for Data Harmonisation

Cynthia A. Huang

cynthia.huang@monash.edu

Department of Econometrics and Business Statistics

supervised by Rob J Hyndman, Sarah Goodwin and Simon Angus

About Me!

👩‍🎓 Final Year PhD Candidate, supervised by, and affiliated with:
- Prof. Rob Hyndman, Econometrics & Business Statistics
  - NUMBATs, Non-Uniform Monash Business Analytics Team
- Prof Simon Angus, Economics
  - SoDa Laboratories, alternative data for social science insights
- Dr. Sarah Goodwin, Human Centered Computing
  - Embodied Visualisation group, immersive technologies for data analysis
💱 Previously: Economics at Unimelb

About Me!

👩‍🎓 Thesis: Unified Statistical Principles and Computational Tools for Data Harmonisation and Provenance
- 🖇️ Data Wrangling and Provenance approaches that incorporate statistical and computational workflow considerations.
- 🌰 Statistically sound, well-documented and low-friction adaptation of “alternative” data for empirical research.
- 🤔 Provenance (approaches, tools and concepts) Across Domains
- 🔀 Unified Framework for recoding and redistributing numeric data between categories

Visualising Category Recoding and Redistributions

Previously presented at VISxPROV Workshop 2023 <arXiv:2308.06535>

Overview

Domain Context

Defining or selecting mappings between classifications or taxonomies,

Task Abstraction

Implementing and validating mappings on given data,

Encoding Design

Documenting and analysing the implemented mapping.

Ex-Post Harmonisation

Ex-post (or retrospective) data harmonization refers to procedures applied to already collected data to improve the comparability and inferential equivalence of measures from different studies (Kołczyńska 2022)