🚀 Project Instructions for mini assessment

Welcome!
This document contains your project tasks and guidelines.
Each participant should complete the task for their selected track only:

• 📊 Data Analyst
• 📈 Data Scientist
• 🤖 Machine Learning Engineer

🕒 Deadline: Submit your project by 26th of June, 2025 Push your work to GitHub and paste your public repo link in the submission form.

---

📊 Data Analyst Project – Supermarket Sales Analysis

📁 Dataset

🔗 Download Dataset (Google Drive)

🎯 Task Objectives

You are to analyze the supermarket sales dataset to uncover trends and key business insights.

1. Data Cleaning

   • Handle nulls and fix data types
   • Convert date/time columns
   • Create new columns (e.g., Total Sales = Unit Price × Quantity)

2. Data Visualization Create at least 3 visuals (Excel / Power BI / Tableau):

   • Top cities / product lines
   • Payment method insights
   • Gender & customer type behavior
   • Sales trend over time
   • Gross Income vs Quantity

3. Insights Summary Briefly answer:

   • What drives the most sales?
   • Underperforming areas?
   • Gender/customer-type behavior differences?
   • Month with highest/lowest sales and why?

4. Push to GitHub Create folder: supermarket-sales-analysis/ Include:

   • Visual dashboard file
   • 2–3 screenshots
   • README.md with summary of findings

---

📈 Data Scientist Project – Loan Default Prediction

📥 Dataset

🔗 Download Dataset (Google Drive)

🎯 Task Objectives

You are to build a predictive model that determines whether a customer is likely to default on their loan based on the dataset provided.

1. Preprocessing & Cleaning

   • Handle missing values
   • Convert categorical features to numeric
   • Remove/impute outliers
   • Create meaningful derived features (optional)

2. Exploratory Data Analysis (EDA) Create at least 2–3 visualizations:

   • Income vs loan default
   • Age group vs loan default
   • Loan grade vs interest rate
     ➕ Explain observed patterns in your notebook or README

3. Model Building

   • Use Logistic Regression, Decision Tree, or Random Forest
   • Evaluate using:
     • Accuracy, Precision, Recall
     • Confusion Matrix
   • Use train/test split or cross-validation

4. Results Explanation Summarize:

   • Key predictive features
   • Business insights from the model

5. Push to GitHub Create folder: loan-default-prediction/
   Include:

   • notebook.ipynb
   • README.md
   • requirements.txt (optional)

---

🤖 Machine Learning Engineer Project – Image Classification

📁 Dataset

🔗 Intel Image Dataset (Kaggle)

Use the Train folder only.
Pick any 3 classes (e.g., forest, mountain, street) Do not use the "Test" or "Predict" folders provided. 👉 Perform your own train/test split in your code.

🎯 Task Objectives

1. Preprocessing

   • Resize images (e.g., 64×64)
   • Normalize pixel values
   • Encode labels

2. Model Building

   • Train a CNN (or use transfer learning: VGG, MobileNet, etc.)
   • Evaluate performance (accuracy or confusion matrix)

3. Streamlit Deployment

   • Build an app that lets users upload an image
   • Show the predicted class

4. Push to GitHub Create folder: intel-image-classification/ Include:

   • notebook.ipynb
   • streamlit_app.py
   • predict.py
   • README.md describing your classes and setup

---

✅ Submission Reminder

• Only submit the GitHub link for your assigned track
• Make sure your repo is public
• Include screenshots or dashboards if needed
• You may add a short video walkthrough (optional)

---

👩🏽‍💻 Good luck! Show us your data skills.