HO CHI MINH UNIVERSITY OF SCIENCE – VIETNAM NATIONAL UNIVERSITY

1.REPUTYZE ASIA - Data Engineer Intern 

Supported developing website and social media data crawlers 

- Developed data crawlers using Selenium Python to fetch posts, user profiles, group profiles from Facebook, Youtube, VOZ.

-Streamlined data workflow by cleaning (remove HTML tags from text), transforming (format raw data to be in-line with the database’s schema) and importing into the company’s system using Python and SQL.

Monitored system performance, processed the DA team requests regarding the integrity of data collection

- Checked for any errors in the system’s data collection process daily and handled them by directly modifying related information in the company’s Database using SQL and Python.

- Automated reporting the status of data collection processes by sending messages over Telegram by using Google Apps Script.

Assisted building an API with Python processing users’ request to detect brands logo in a picture or video

- Trained a custom YOLOv5 model for object detection. Went through the phases of data collection (2000 images), preprocessing (noise removal, annotation, augmentation), training, and model evaluation.

- Integrated pre-trained model modules, such as Object Segmentation (MRCNN), Character Recognition (OCR - EasyOCR) to the API.

- Programmed a data ingestion module transporting log files, results from the models to Elasticsearch via Logstash using Python.

- Designed Kibana dashboards to visualize the model results and performance indicators to gain insights into the visibility of brands in media content and evaluate the model's effectiveness.

2. DATA ANALYSIS OF F&B ESTABLISHMENTS IN HO CHI MINH CITY - Class Project

- Used Selenium Python to collect users’ comments and review data from more than 80.000 restaurants on Foody website.

- Performed exploratory data analysis (EDA) and discovered 6 descriptive insights about price, ratings and customers’ preferences regarding restaurants’ services.

- Collected about 2000 comments from Phuc Long and Highlands Coffee stores in HCMC to analyze the differences in price, space, location, and reviews between stores of each brand.

- Categorized customers’ sentiment based on high/low rating stores and positive/negative comments using keyword extraction and sentiment analysis.

3. DANNY MA – 8 WEEKS SQL CHALLENGE - Personal Projec 

Completed 8 case studies using SQL queries. •

The challenges involved cleaning data, using JOIN techniques, sub-query, common table expression (CTE) and writing functions to answer pseudo-business questions.