Denis Ndemwa

Inventing at the intersection of circuits, code, and intelligent systems.

🧪 Lab Challenges

A curated set of lab challenges and mini-projects that blend hands-on practice, critical thinking, and applied data science techniques.
These labs come from self-study, coursework, and personal curiosity.

📊 Business Intelligence with Power BI

Problem Statement

Build an interactive dashboard to analyze and visualize product sales, regional performance, and key business metrics for a fictional retail company.

Approach

Imported sales data into Power BI.

Performed data cleaning and transformation using Power Query.

Created measures using DAX for KPIs like Total Sales, Average Profit Margin, and Year-over-Year Growth.

Designed an interactive report with slicers and drill-through pages to analyze trends by product category and region.

Tools Used

Microsoft Power BI

Power Query

DAX

🖼️ Screenshots

Loading data

bi load

Creating relationships

bi l

Transforming Data

bi t

Different Dashboards

bi

Dashboard 2 bi 2

Dashboard 3 bi 1

Key Lessons Learned

Efficient use of Power Query for data wrangling and combining multiple data sources.

Writing reusable DAX measures to keep calculations clean and scalable.

Importance of designing user-friendly and intuitive dashboards for non-technical stakeholders.

Data Visualization using Tableau

Problem Statement

Create an HR analytics dashboard to visualize workforce composition, turnover trends, and diversity metrics.

Approach

Connected Tableau to an HR dataset in Excel.

Cleaned and reshaped data to prepare for analysis.

Built multiple visualizations: bar charts for turnover, pie charts for diversity, and trend lines for hiring vs. attrition over time.

Combined visualizations into a single interactive dashboard with filters and tooltips.

Tools Used

Tableau Desktop

Tableau Prep (optional, for data cleaning)

Excel

🖼️ Screenshots

Building data source from a ChatGPT prompt

image

  1. Connect data

image

3. data visualization

image

** 4. Creating test field** image

5. Creating BANNS

image

** 6. Creating Calculated fields**

image

** 7. Creating Charts**

image

** 8. Creating dashboard by importing and positioning all charts to the dashboard using text to label the dashboard, horizontal and vertical containers to house all the charts and dividers to separate charts**

image

Dashboard 1 image

Dashboard 2 image

Key Lessons Learned

Designing dashboards with clear storytelling and logical flow.

Using calculated fields and level-of-detail (LOD) expressions to create advanced metrics.

Fine-tuning tooltips, colors, and formatting for better usability and visual appeal.

🌐 Web Scraping

Problem Statement:
Extract structured data from a website to analyze sports statistics.

Approach:

  • Identified HTML structure and data elements.
  • Used BeautifulSoup for parsing static pages.
  • Stored data in CSV files for analysis.

Tools Used:
Python, BeautifulSoup, pandas

Key Lessons Learned:

  • How to parse HTML effectively.
  • Cleaning and structuring raw scraped data.
  • Respecting robots.txt and ethical scraping practices.

🎬 Netflix Data Wrangling

Problem Statement:
Prepare Netflix title data for analysis by handling inconsistencies and missing values.

Approach:

  • Explored dataset to understand schema and data quality.
  • Handled missing data, standardized date formats, and normalized categorical fields.

Tools Used:
Python, pandas

Key Lessons Learned:

  • Importance of initial data profiling.
  • Data cleaning techniques: fillna, dropna, and string manipulation.
  • Preparing data to support later visualization and modeling.

📊 Exploratory Data Analysis (EDA)

Problem Statement:
Identify trends and insights in the cleaned Netflix dataset.

Approach:

  • Created visualizations for distributions, correlations, and trends.
  • Highlighted content trends over time and popular genres.

Tools Used:
Python, matplotlib, seaborn, pandas

Key Lessons Learned:

  • Visual storytelling: turning raw data into insight.
  • Choosing the right plot types for different questions.
  • Spotting outliers and data quality issues through visuals.

🧠 Interview with Geoffrey Everest Hinton (Godfather of AI)

Problem Statement:
Summarize expert insights on deep learning, AI ethics, and the future of AI.

Approach:

  • Watched the interview, created structured notes.
  • Highlighted quotes and reflected on implications for machine learning.

Tools Used:
Markdown, note-taking apps (Notion, Obsidian)

Key Lessons Learned:

  • Value of learning directly from pioneers.
  • Importance of critical thinking in AI.
  • Ethical considerations as AI evolves.

📈 Regression Models

Problem Statement:
Predict numeric outcomes (e.g., ratings or views) using regression.

Approach:

  • Engineered features.
  • Applied Linear Regression and regularized models (Ridge, Lasso).
  • Evaluated using RMSE and R².

Tools Used:
Python, scikit-learn, pandas

Key Lessons Learned:

  • Impact of feature selection and scaling.
  • Benefits of regularization to reduce overfitting.
  • Using metrics to compare model performance.

🧪 Classification Models

Problem Statement:
Classify items into categories, like “popular” vs “niche” titles.

Approach:

  • Prepared categorical and numerical features.
  • Applied Logistic Regression and Decision Tree classifiers.
  • Evaluated accuracy and visualized confusion matrix.

Tools Used:
Python, scikit-learn, pandas

Key Lessons Learned:

  • Handling class imbalance.
  • Trade-offs between model simplicity and accuracy.
  • Choosing the right evaluation metrics for the task.

🔧 MLOps

Problem Statement:
Build and deploy an end-to-end machine learning pipeline.

Approach:

  • Structured project with modular scripts.
  • Tracked experiments using MLflow.
  • Explored Docker for deployment and reproducibility.

Tools Used:
Python, MLflow, Docker, scikit-learn

Key Lessons Learned:

  • Reproducibility in ML workflows.
  • Tracking and comparing experiments.
  • Basics of CI/CD and deployment readiness.

Field Lab: Supervising a Construction Site

Problem Statement

Oversee daily operations at an active construction site to ensure safety, quality, and project milestones are achieved — all while managing people, equipment, and unforeseen challenges.

Approach

Daily planning: Started each morning by reviewing schedules, checking weather forecasts, and briefing teams.

Site inspection: Conducted walk-throughs to verify safety compliance (PPE use, secured scaffolding, clean pathways).

Coordination: Communicated with subcontractors, suppliers, and engineers to align on daily targets.

Problem-solving: Resolved on-site issues like delayed material deliveries, equipment breakdowns, and design clarifications.

Reporting: Logged daily progress reports, incidents, and resource usage.

Tools Used

Construction management software (e.g., Procore, Buildertrend)

Mobile apps for site reporting and checklists

Standard safety equipment: helmet, vest, boots

Radio and phone for quick team coordination

#Key Lessons Learned

Proactive communication is vital — delays are often prevented by clarifying plans early.

Safety isn’t a checklist — it’s part of daily habits and team culture.

Flexibility and quick thinking help handle surprises like weather, late deliveries, or design changes.

Clear documentation helps protect the team and keep stakeholders aligned.

“Transforming vision into built reality with precision and technology.”