🔧 Project: Predict Titanic Survivors using Microsoft Fabric

🚀 Project Overview

In this beginner-friendly machine learning project, we're using Microsoft Fabric to analyze the Titanic passenger dataset and build a predictive model that determines whether a passenger survived the disaster.

We’ll go step-by-step using Lakehouse, Notebooks, and PySpark (Python) — all inside Microsoft Fabric’s unified data platform.

✅ Lakehouse

A Lakehouse in Azure Fabric gives you a single platform to store, process, and analyze both raw and structured data, making it ideal for modern data science and analytics projects.

✅ Traditional Approach (Before Lakehouse)

➡️ Problem: These were separate systems. You had to move data from the lake to the warehouse for reporting, which caused duplication, delay, and complexity.

✅ Unified Storage in Microsoft Fabric (Lakehouse)

Now with OneLake in Microsoft Fabric:

• You store all your data in one place — no need to move between lake and warehouse.

• The same data can be:

  • Accessed as files (CSV, Parquet) → for ML/AI

  • Queried as structured tables → for Power BI/SQL

🧠 Example:

Imagine you upload titanic.csv to the Lakehouse:

• In Files: It exists as a raw file (data lake style)

• When you Load to Table: It becomes a structured table (warehouse style)

• Power BI, SQL, and notebooks all use the same storage location (OneLake)

Files

Tables

🔄 Summary:

“Unified storage” means you don’t need separate systems for storage (lake) and analytics (warehouse). With OneLake, your data serves both data science and business intelligence directly — faster, simpler, and cheaper.

1. Create a new Lakehouse

• Go to My Workspace

• Click + New item → Lakehouse

• Name it Titanic_Lakehouse

2. Upload the dataset

• Download titanic.csv from GitHub link
  https://raw.githubusercontent.com/datasciencedojo/datasets/master/titanic.csv

• In your Lakehouse, go to the Files tab

• Click Upload and select titanic.csv

3. Convert the file into a structured table

• After the file uploads, right-click on titanic.csv

• Choose Load to Tables → New table

• Name the new table titanic and click Load

4. Open a Notebook

• At the top menu, click Open notebook → New notebook

• Make sure the notebook is attached to Titanic_Lakehouse

• Read the dataset using PySpark

  • Run this code:

      df = spark.read.table('titanic')
      display(df.limit(5))
    

    

• Print Schema

  

EndFragment