Completed Certification Course on Applied Data Science Lab Course by WorldQuant University

I have completed eight projects. Each project consists of four self-paced lessons, followed by an assignment that is programmatically graded. For each assessment,need to score 90% or better to move further project.

1. Housing in Mexico: — Data Analysis Project

This project uses a dataset of 21,000 properties in Mexico to investigate whether property price is more strongly associated with property size or location.

Learning objectives
Learners will:

1. Import data
2. Clean the data
3. Explore the data
4. Create visualizations
5. Measure correlation
6. Investigate location

 Expected conclusion: The analysis should not assume in advance that either factor is more important. The correlation analysis measures the linear association between size and price, while the location analysis reveals how prices differ between areas. Together, these provide evidence for assessing the relative role of property size and location.

 

2. Apartment Sales in Buenos Aires — Project Overview

This project focuses on predicting apartment prices in Buenos Aires, Argentina, using linear regression.

Main learning objectives

1. Load and explore the data
2. Build a data preprocessing pipeline
3. Train a linear regression model
4. Evaluate model performance
5. Reduce overfitting

Typical machine-learning workflow:

CSV data → Data cleaning → Train/test split → Preprocessing pipeline → Linear/Ridge regression → Evaluation → Reduce overfitting → Final predictions

Expected outcome:

By the end of the project, learners should be able to build an end-to-end regression pipeline that takes raw apartment data and produces predicted property prices, while handling missing values and categorical variables and controlling overfitting.

If this is for a Data Science/ML course assignment, the core concepts demonstrated are data preprocessing, pipelines, categorical encoding, linear regression, model evaluation, and regularization.

3. Air Quality in Nairobi — Project Overview

This project focuses on time-series forecasting of particulate matter (PM) levels in Nairobi, Kenya.

Learners will:

• Extract air-quality data from MongoDB using pymongo.
• Clean and prepare the time-series data, including handling missing values and selecting the relevant particulate-matter measurements.
• Explore temporal patterns such as trends, seasonality, and autocorrelation.
• Build an ARMA (AutoRegressive Moving Average) model to forecast future particulate-matter levels.
• Tune the ARMA hyperparameters (AR and MA orders) to improve forecasting performance.
• Evaluate the model using appropriate time-series metrics and compare predictions with observed PM levels.
• Interpret the results to understand how historical air-quality measurements can help predict future pollution levels.

Key technologies: Python, pymongo, MongoDB, pandas, NumPy, statsmodels, and time-series forecasting.

 

 4. Earthquake damage in Nepal

This project focuses on using machine learning to predict the level of earthquake damage to buildings in Nepal.

What learners do:

1. Extract data from SQLite
2. Clean and prepare the data
3. Build machine-learning models
4. Evaluate model performance
5. Investigate data bias and discrimination

Key learning outcomes

By completing this project, learners practice:

SQLite → Data extraction → Data cleaning → Feature engineering → Logistic Regression + Decision Tree → Model evaluation → Bias/fairness analysis

The important lesson is that a model can appear accurate while still producing unfair results if the underlying training data contains systematic biases.

 

Certification Badge by Credly.com . You can view more details from badge.

 

5. Bankruptcy in Poland:

This project focuses on predicting whether a company in Poland will go bankrupt using machine learning.
What learners do

1. Explore the dataset
2. Use Linux command line
3. Handle imbalanced data
4. Build machine-learning models
5. Evaluate with precision and recall
6. Compare model performance

 Key learning outcome:

The project teaches learners that machine-learning performance is not just about accuracy. For an imbalanced problem such as bankruptcy prediction, understanding precision, recall, resampling, and model behavior is essential for interpreting whether a model is genuinely useful.

 

6. Customer Segmentation in the US:

This project focuses on unsupervised machine learning to identify groups of US consumers with similar characteristics.

Key learning objectives:

• Clean and prepare consumer/customer data.
• Apply K-Means clustering to divide customers into meaningful segments.
• Determine an appropriate number of clusters, commonly using the Elbow Method or silhouette analysis.
• Use Principal Component Analysis (PCA) to reduce high-dimensional data to 2D or 3D for visualization.
• Interpret the characteristics of each customer segment.
• Build an interactive Plotly Dash dashboard to explore clusters and customer characteristics.
• Develop practical skills in Python, scikit-learn, PCA, K-Means, Plotly, and Dash.

Typical workflow:

Customer Data → Data Cleaning → Feature Scaling → K-Means → PCA → Cluster Visualization → Plotly Dash Dashboard

Main concepts:
K-Means groups customers based on similarity, while PCA transforms multiple variables into fewer dimensions that make the clusters easier to visualize..

7. A/B Testing at WorldQuant University

In this project, learners investigate whether sending an email increases enrollment in a program at WorldQuant University (WQU).

Key skills covered:
A/B testing: Compare two groups:

• Control group: Did not receive the email.
• Treatment group: Received the email.

Chi-square test: Determine whether email receipt and program enrollment are statistically associated.

Hypothesis testing:

• Null hypothesis (H₀): Email sending and enrollment are independent.
• Alternative hypothesis (H₁): Email sending and enrollment are associated.

Python: Build custom classes to implement an ETL (Extract, Transform, Load) workflow.
Data application: Develop an interactive application using a three-tier architecture:

• Presentation layer – user interface
• Application/business layer – processing and logic
• Data layer – data storage and retrieval

The project combines statistical inference, Python OOP, ETL, and application architecture in a practical A/B-testing scenario.

 

8. Volatility Forecasting in India:

Learners build a GARCH (Generalized Autoregressive Conditional Heteroskedasticity) time-series model to forecast asset volatility in the Indian financial market. They obtain historical stock-market data through an API, clean and organize the data, and store it in a SQLite database. Finally, they develop their own API to serve the model’s volatility predictions, creating an end-to-end workflow from data acquisition to deployment.

Skills Gained from these modules in summary: 

• API Design
• Data Science
• Data Visualization
• Machine Learning
• MongoDb
• Python (Programming Language)
• SQL
• Statistics.

You can Pursue the course here: https://www.wqu.edu/adsl