Skip to content
View avisionh's full-sized avatar
🧽
🧽

Block or report avisionh

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
avisionh/README.md

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:

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


Connect with me:

avisionh | Medium avisionh | LinkedIn avisionh | Instagram



Languages and Tools:

aws

gcp

terminal

sqlite

tsql

neo4j

R

RStudio

Python

Jupyter

PyCharm

Git

GitHub

AzureDevOps

travisci

markdown

LaTeX

Slack

Trello

Pinned Loading

  1. alphagov/govuk-entity-personalisation alphagov/govuk-entity-personalisation Public archive

    Using entities from NER on GOV.UK content to power personalisation.

    Jupyter Notebook 3 4

  2. Hong-Kong-Districts-Info/dashboard-hkdistrictcouncillors Hong-Kong-Districts-Info/dashboard-hkdistrictcouncillors Public

    Shiny app dashboard of HK district councillors' information include FB pages.

    R 6 3

  3. camdencrime camdencrime Public

    Visualising Camden crime data and presented on web-hosted Jupyterbook.

    Jupyter Notebook 1

  4. preferenceallocation preferenceallocation Public

    Bespoke algorithm for tackling one-sided matching problem where we have preferences to consider.

    R 1 1

  5. sqlquerygraph sqlquerygraph Public

    Represent your SQL queries as a graph.

    Python 3

  6. sqlgpt sqlgpt Public

    Translating natural language into a query and then executing it on a database

    Python 2