Camille Sze Pui Ko

Logo

Volunteer Data Analyst @ Infoxchange | PT Lecturer @ School of CS, Adelaide U

My Skills AWS Certified Cloud Practitioner (CLF-C01) Microsoft Certified Azure AI Engineer Associate (AZ-102) Microsoft Certified Azure Power BI Data Analyst Associate (PL-300) Microsoft Certified Azure Data Fundamentals (DP-100) Microsoft Certified Azure AI Fundamentals (AI-900) Tableau Certified Data Scientist Tableau Certified Data Analyst
camillekokoko
View My LinkedIn Profile

View My GitHub Profile

breadcrumb > You are viewing : Portfolio Click me to Conference-Community-Hackathon
Navigate To Anchor:
AI Engineering Machine Learning and Deep Learning DevOps/MLOps/Cloud
GIS Statistical Analysis Programming Language Proficiency

Data Analytics and Visualisation

1. Reporting for 2021-2023 on Community Services within the City of Whyalla

Python PowerBI GA-UA CRM-Database Canva Excel

As a data analyst at SA Community, a directory supported by the Government of South Australia to maintain updated information on community services across the state, my role involves analyzing and presenting council-based analytics summaries related to community services and information demands. This analytical work provides valuable insights into the social and demographic needs of a council. This information empowers informed decision-making by identifying service gaps, ensuring equitable distribution, and facilitating the effective sharing of community grants, resources, and assets. The result is an enhancement in support systems and overall outcomes for the entire community.

PowerBI Whyalla 2022-2023

View the PowerBI dashboard of Whyalla 2022-2023 at this external link
View my experience at Infoxchange

2. E-commerce Optimization - AB Testing

BigQuery GitLab E-commerce Tracking AB Testing SQL Hotjar

Spearheaded the enhancement of online sales and user experience as a data analyst/statistician, managing GA4 implementation, AB testing, and data analytics, resulting in a remarkable ~15% increase in conversion rates. Additionally, implemented sophisticated e-commerce tracking to meticulously monitor and optimize product performance.

Visualization of Conversion Funnel AB Testing Experimentataion Visualization of Conversion Funnel AB Testing Experimentataion

View the pdf of one of the winner analysis on rearranging tabs product anchor experiment

3. User Journey and Engagement Study

BigQuery Google Cloud Platform Amazon Web Service E-commerce Tracking Python data studio

As the sole analyst and engineer, I independently implemented GA4 and conducted data modelling, extract-load-tranform (ELT) and metrics such as exit rate, view item rate, contributing % in new users, bounce rate, etc. to understand user behavior and enhance content engagement. Achievements include the proficient utilization of custom dimensions for detailed product/web analytics and a 20% improvement in engagement metrics through data-driven UI/UX adjustments. My expertise spans GA4 implementation, event tracking, audience segmentation, data warehouse management, and integration with platforms like Google BigQuery.

Combine_ga4_bigquery_for insight

Check out the code here

4. SEO

tableau social listening SEO excel Python data studio

Throughout my role as a Reporting and Data Analyst, I excelled in optimizing SEO strategies and producing automated daily reports to meticulously measure and enhance overall performance. Achieved notable improvements in website visibility, engagement, and key performance indicators, contributing to data-driven decision-making processes.

SEO

AI Engineering

1. Leveraging large language model with domain adaptation: enhancing community directories with deep learning text summarization for machine-readable cataloging (MARC) standard description notes

BERT PyTorch LLM Git VSCode Docker Java

This research project, conducted as part of my Master’s program in AI and Machine Learning at the University of Adelaide, explores the integration of state-of-the-art technologies such as BERT, PyTorch, and Large Language Models (LLM), domain adaptation in the field of AI engineering and Natural Language Processing (NLP). The focus is on optimizing community directories, particularly utilizing the SAcommunity open data database with around 14,500 records, by generating MARC standard description notes for the new summary field in the directory using deep learning-based text summarization.

Python and PyTorch served as the primary languages, complemented by SQL for database integration and Java for specific functionalities. Utilizing a streamlined software engineering workflow, I employed Git for version control, Docker for containerization, and database integration to manage data effectively. The project included building a seamless pipeline for data extraction and deploying the model to the web, showcasing the intersection of AI engineering and software development.

View exBERTSum’s Code on GitHub
View exBERT’s Code on GitHub

Machine Learning and Deep Learning

1. Optimizing User Personas in Newspapers: A Machine Learning Approach with K-means and Random Forest Models

BigQuery Google Cloud Platform E-commerce Tracking Machine Learning Python

As a Data Analyst, I skillfully employed both claimed and observed data to derive valuable insights and inform decision-making processes in marketing, advertising, content planning, and other related domains. This approach ensured a comprehensive understanding of user behavior and preferences for effective strategy implementation.

persona rf kmean

2. Explainable AI with Shapley values

View code here

3. Image Classification

This project aims to investigate image classification using the CIFAR-10 dataset, employing two popular deep-learning architectures: ResNet50 and VGG16.

Check out code

4. Deep learning

Check out code

5. RNN

Check out code

6. Precision Small Object Detection in Real-Life Applications using Fasterrcnn_resnet50_fpn and SSD300_vgg16 Algorithms

PyTorch Computer Vision Anaconda Python

This is one of the computer vision I have done during my research and pursuit of a master’s degree in AI and machine learning. The project utilized algorithms with a primary focus on enhancing precision in detecting small objects within real-world contexts. The implementation showcased a robust and efficient solution, achieving high accuracy in object detection for improved practical applications. Notable features include the integration of the ResNet50 backbone network and Feature Pyramid Network (FPN) in Fasterrcnn_resnet50_fpn for efficient and accurate detection. Additionally, SSD300_vgg16 employed a Single Shot Multibox Detector (SSD) with a VGG16 backbone, ensuring real-time object detection across diverse scenarios.

Check out the code

7. Identify Playtus - Endangered Animal - using Computer Vision

playtus

View pdf

8. Segment Computer Vision

mmseg

View code

DevOps/MLOps/Cloud/Data Engineering

1.Database

2.Data Structures and Algorithms

3.Cloud Engineering on AWS EC2 and Docker

architecture: architecture

View here

4.Airflow

5.Snowflake

Snowflake credit: snowflake

Check out code

Geographic Information System (GIS)

Statistical Analysis

summary View Code here
View Code here

stat View Code here

Programming Language Proficiency

1. C# Programming C#

2. Java Programming Java

3. C Programming C

4. Django - python web framework Django Python

Navigate To Anchor:    
Statistical Analysis Geographic Information System (GIS) DevOps/MLOps/Cloud
Machine Learning and Deep Learning AI Engineering Data Analytics and Visualization

breadcrumb > You are viewing : Portfolio | Click me to Conference-Community-Hackathon

Page template forked from evanca