JoliePhamPortfolio

Portfolio that consists of my personal data analysis / data science projects

Project 1: City of Melbourne

• The aim is to find out which suburb is the most liveable in City of Melbourne, based on property prices, facilities and overall safety perceptions.
• Located and merged data from 3 different sources and tidied the data using R and Tidyverse.
• Scanned for and dealt with all missing values, inconsistencies and outliers.
• Transformed data for better understanding and help decreasing the skewness in distribution.
• Visualised and communicated the results with ggplot2

Project 2: Spam Detection

• The aim is to classify text messages (sms) as either spam or ham, based on the content
• Prepared and cleaned raw data, conducted EDA in Python using Pandas and Matplotlib to form an interpretation of data.
• Tokenised the spam/ham sms by words using CountVectorizer.
• Trained and evaluated KNN and Decision Tree using SMOTE

Project 3: Employee Turnover Analysis

• The aim is to create a model that can predict groups of employees that have a high tendency of leaving company.
• Explored the data and relationships using descriptive analytics and graphs using Python.
• Transformed, normalised data to prepare for training models, using K-Means and DBSCAN
• Tested models and evaluated clustering performance.

Project 4: Recommendation Engine Algorithms Evaluation

• The aim is to evaluate the company’s new recommendation engine algorithm is worth rolling to all customers.
• Performed hypothesis testing to calculate the significance of key results on samples.
• Identify trends by performing linear and multiple regression analysis.
• Created visualisations to support communication using ggplot2 and plotly packages in R.
• Provided recommendations for future data collection and investigation to minimise bias and errors.