Hi, my name is Yaohong Liang
I'm a Data Analytics professional!

The dashboard summarizes unicorns across industries, highlights the investors backing them, and presents their valuations, providing a clear overview of the broader unicorn ecosystem.

Tools:

• Google Data Studio (Looker)
• Excel

Skills:

• Data Wrangling
• Exploratory Data Analysis
• Data Visualization

The analysis will cover crucial areas such as revenue and retention trends, cohort behavior, the impact of discounts on pricing, and customer segmentation. Additionally, more advanced techniques, including lifetime value (LTV) forecasting, cross-selling analysis will be employed to provide deeper insights into customer behavior and business performance.

Tools:

• Python

Skills:

• Revenue Analysis
• Retention Analysis
• Discount Impact Analysis
• Customer Segmentation

In this project, I performed exploratory analysis to understand the fraud transactions on a bank simulated dataset, and I also applied statistical methods to detect fraudulent activities.

Tools:

• Python

Skills:

• Exploratory Data Analysis
• Data Sampling
• Statistical Modeling

To understand the development of the electric vehicle market, this project studies the data of electric vehicles registered in WA from 1997 to 2023.

Tools:

• R

Skills:

• Market Research
• Team Collaboration
• Exploratory Data Analysis
• Data Visualization

In this project, I performed analysis on customers' reviews on an E-commerce site to identify their areas of interest/concern.

Tools:

• Python

Skills:

• Natural Language Processing
• Topic Modeling
• Keywords Analysis

I developed supervised learning algorithms for customer churn prediction in this project. The labelled data in this data set is imbalanced, so I applied SMOTE for oversampling. Besides, I applied encoding, standardization technique to transform the features. Logistic Regressions, KNN, Random Forest algorithms are used for modeling. Model evaluation involves metircs like f1-score, ROC and AUC scores.

Tools:

• Python

Skills:

• Data Pipeline
• Predictive Modeling
• Model Evaluation

In this project, I used machine learning techniques to build models that can detect fraud credit card transactions on a highly imbalanced dataset, in which only less than 1% transactions are considered fraud. Random downsampling method is used to handle the imbalance data.

Tools:

• Python

Skills:

• Machine Learning
• Classification Modeling

This analysis aims for developing a statistical model to classify heart disease using data collected through non-invasive procedure. The final model achieves 84% accuracy and has a false positive rate of 18%.

Tools:

• R

Skills:

• Statistical Modeling