Presentation Script: Business Intelligence Implementation for Al Naseej Al Sorayi Carpeting Company

Introduction

Good morning/afternoon everyone. Today, I'm excited to share with you our journey of implementing a Business Intelligence solution for Al Naseej Al Sorayi Carpeting Company. This presentation will walk you through our evolution from a simple Excel-based approach to a robust SQL Server implementation, highlighting the challenges we faced, the solutions we developed, and the significant improvements we achieved.

Company Background (1 minute)

Al Naseej Al Sorayi is a leading carpet manufacturing company based in Jeddah, Saudi Arabia. Established in 1985, they've grown to become one of the largest carpet producers in the Middle East with three main business units:

• Tufting Division: Producing residential carpets
• Weaving Division: Creating high-end residential and commercial carpets
• Yarn Division: Manufacturing yarn for internal use and external sales

The company was facing several challenges with their reporting processes, including manual data collection, inconsistent data across departments, and time-consuming consolidation of production and inventory information.

Project Overview (2 minutes)

Our project had several key objectives:

• Centralize data from multiple Excel sources
• Create interactive dashboards for production monitoring
• Implement automated data refresh processes
• Enable drill-down analysis capabilities
• Improve decision-making with real-time insights

We approached this project in two distinct phases:

1. Initial implementation with direct Excel import to Power BI (Summer 2024)
2. Redesigned implementation with SQL Server after completing a Business Intelligence course (Spring 2025)

Let me show you the bird's eye view of the three dashboards we created to monitor production and inventory.

[SHOW BIRD'S EYE VIEW SLIDE]

These dashboards provide comprehensive monitoring of:

• Monthly production performance
• Daily production trends
• Inventory values across business units

Initial Approach: Excel to Power BI (3 minutes)

In our initial approach, we connected Power BI directly to Excel files. Let me walk you through this process:

[SHOW INITIAL APPROACH SLIDE]

The workflow was straightforward:

1. Collect Excel files from different departments
2. Clean and format data in Excel
3. Import Excel files directly into Power BI
4. Create relationships in Power BI data model
5. Build visualizations and dashboards
6. Manually refresh data when Excel files were updated

For guidance, we relied heavily on Microsoft's documentation, particularly "From Excel workbook to stunning report" which provided step-by-step instructions for transforming Excel data into Power BI reports.

[SHOW EXCEL CONNECTION SCREENSHOT]

While this approach allowed us to quickly create dashboards, we soon encountered several challenges:

• Performance issues with large datasets
• Limited data modeling capabilities
• Manual update processes that were time-consuming
• Data inconsistencies across departments
• Limited security options
• Scalability constraints due to Excel file size limitations

Let me show you the first version of our dashboards that were created using this Excel-based approach.

[SHOW ORIGINAL DASHBOARDS]

As you can see, we created three main dashboards:

1. Production Dashboard with monthly KPIs and utilization metrics
2. Production Flow Dashboard showing daily production trends
3. Inventory Dashboard displaying inventory values by period

These dashboards provided value by consolidating data and creating interactive visualizations, but the underlying architecture had significant limitations.

Business Intelligence Course: The Catalyst for Change (2 minutes)

In March 2025, I enrolled in a comprehensive Business Intelligence course that completely changed our approach. This course covered:

• Data warehouse concepts and best practices
• Star schema and snowflake schema design
• Dimension and fact table modeling
• ETL processes and data transformation
• SQL Server Management Studio usage
• Advanced Power BI data modeling techniques

The course revealed that our Excel-based approach was not optimal for enterprise BI implementation. I learned that proper data modeling could significantly improve performance, and SQL Server would provide better security and scalability.

Improved Approach with SQL Server (5 minutes)

Based on the knowledge gained from the course, we redesigned our BI implementation using SQL Server as the foundation.

[SHOW IMPROVED APPROACH FLOW DIAGRAM]

Our new approach consisted of three main phases:

1. Data Collection: We still collected the same Excel files, but instead of connecting Power BI directly to them, we used them as source data for our SQL Server database.

2. BI Process: This was the critical new phase that was missing in our original approach.

   • We created a SQL Server database
   • Implemented a star schema design
   • Developed an ETL process
   • Transformed the data to optimize for analysis
3. Dashboard Creation: We connected Power BI to our SQL Server database, optimized the data model, and recreated our dashboards.

Let me show you the ETL process we implemented:

[SHOW ETL DIAGRAM]

The ETL process consists of three key steps:

• Extract: We used SQL Server Import Wizard to extract data from Excel files into staging tables
• Transform: We applied data cleaning, standardization, and business rules using SQL queries
• Load: We loaded the transformed data into our star schema database

Let me demonstrate how we used SQL Server Management Studio for this implementation:

[SHOW SQL IMPORT SETUP SCREENSHOT]

Here you can see the SQL Server Import Wizard configuration screen where we selected Excel as the data source and mapped columns to our staging tables.

[SHOW SQL IMPORT PROGRESS SCREENSHOT]

This screen shows the progress of data transfer with row counts for each table, providing real-time feedback on the import process.

Star Schema Implementation (3 minutes)

The star schema design was a critical component of our improved approach:

[SHOW STAR SCHEMA DIAGRAM]

Our star schema consists of:

• Dimension Tables:

  • DimDate: Contains date-related attributes
  • DimBusinessUnit: Contains business unit information
  • DimProductType: Contains product type information
  • DimWarehouse: Contains warehouse information

• Fact Tables:

  • FactProduction: Contains production metrics
  • FactInventory: Contains inventory metrics
  • FactKPI: Contains KPI metrics

This design significantly improved query performance and simplified complex business relationships.

Power BI Connection to SQL Server (2 minutes)

Once our star schema was created, we connected Power BI to the SQL Server database:

[SHOW POWER BI NAVIGATOR SCREENSHOT]

Here you can see Power BI's Navigator interface showing our SQL Server database structure with dimension and fact tables. This connection method provided several advantages:

• Improved performance with optimized queries
• Enhanced data modeling capabilities
• Automated refresh capabilities
• Better security integration

Now, here's the key innovation in our approach: Instead of recreating all our visualizations from scratch, we created database views in SQL Server that matched the structure of our original Excel files. This allowed us to simply change the data source in our existing Power BI reports without having to rebuild all the visualizations.

This approach saved us significant time and effort while still gaining all the benefits of the SQL Server implementation.

Improved Dashboards (3 minutes)

Let me show you the improved versions of our dashboards:

[SHOW IMPROVED DASHBOARDS]

While they look similar to our original dashboards, the underlying architecture is completely different, resulting in:

• Faster loading times
• More responsive filtering
• Automated data refresh
• Improved security
• Better scalability

Let's look at the performance improvements:

• Dashboard load time reduced from 12-15 seconds to 2-3 seconds (80-85% faster)
• Data refresh time reduced from 8-10 minutes to 2-3 minutes (70-75% faster)
• Filter response time reduced from 3-5 seconds to less than 1 second (80-90% faster)
• Power BI file size reduced from 45 MB to 18 MB (60% smaller)

Business Benefits (2 minutes)

The improved BI implementation delivered significant business benefits:

Operational Benefits:

• Improved decision-making with real-time data access
• Operational efficiency through daily production tracking
• Enhanced inventory management
• Increased accountability with clear performance metrics
• Time savings of approximately 15 hours per month in manual data preparation

Technical Benefits:

• Centralized data repository providing a single source of truth
• Enhanced security with database-level permissions
• Improved data quality with validation rules
• Scalability to accommodate business growth
• Better maintainability for updates and changes

Lessons Learned and Conclusion (2 minutes)

This project taught us several valuable lessons:

1. Proper data modeling is critical for BI success
2. SQL Server provides significant performance advantages
3. Star schema simplifies complex business relationships
4. SQL Import Wizard streamlines data loading
5. Business Intelligence education is invaluable for effective implementation

In conclusion, the Business Intelligence course in March 2025 was the catalyst that transformed a simple Power BI project into a robust, enterprise-grade BI solution. By implementing a proper star schema database in SQL Server, we achieved significant improvements in performance, scalability, and maintainability.

The project demonstrates the importance of proper data modeling and database design in Business Intelligence implementations, and shows how even small changes in approach can lead to substantial improvements in outcomes.

Thank you for your attention. I'm happy to answer any questions you may have.

I'll create alternative star schema designs for each fact table following the traditional approach, where each fact table connects to at least three dimension tables.

Key Improvements in the Revised Star Schema Design:

1. More Dimensional Context: Each fact table now connects to at least four dimension tables, providing richer analysis capabilities.

2. Enhanced Granularity: The redesigned schema allows for more detailed analysis by:

   • Adding DimPlant to FactProduction for location-based analysis
   • Adding a dedicated DimComparisonType table for FactProductionComparison
   • Expanding dimension tables with additional attributes
3. Better Performance: With proper indexing on foreign keys, this design would support faster query execution and easier filtering.
4. Standardized Dimensions: The design uses consistent dimension tables across fact tables, promoting reusability and simplified maintenance.

5. SQL Implementation Considerations:

   • Each dimension table should have clustered indexes on primary keys
   • Foreign keys in fact tables should have non-clustered indexes
   • Dimension tables should include effective dates if slowly changing dimensions are needed

This traditional star schema design provides better analytical capabilities while maintaining the simplicity that makes star schemas effective for business intelligence applications.