avisionh / predictgraduateprospects

Applies Bayesian techniques for analysing various factors that can influence a UK university's graduate prospects rating from the HESA SFR247 and Complete University Guide table.

Hi there 👋

Working on: Writing a mini-'package' to extract table names from SQL scripts and visualise in a graph database. Available here. Creating a network graph of media content. Available here. Open-source package contributions: Interactive bokeh visualisation of SOMs in the minisom package. travis CI instructions to use DPL-V2 in the mdBook package. R functionality for pre-commit hooks to govcookiecutter package.

Highlights: Used the custom-trained NER outputs to make GOV.UK's search engine more explainable, transparent and powerful through CBOW and BERT word embeddings to get content-specific synonyms which can enrich the existing entities we have on our content. Available here. Utilised doc2vec, USE, ANNOY and SOMS to identify semantically-similar documents on GOV.UK. Available here. Applied object-orientated programming principles to identify the language of content. It is available here. Built a website/Shiny app that empowers users to learn more about their local district councillors in Hong Kong. Let me know if you're interested! Repo is here. 🔭 Modelled the causal drivers behind survey responses on longitudinal data conducted on pupils during COVID-19. It is available here. Searching for: Interesting, open-source projects to collaborate on - get in touch! 👯 Public talks EARL 2019 Conference Speaker Interview for Data Strategy video series Interview for Martin Chan's blog

---

Connect with me:

---

Languages and Tools: