onyangoju/3-Tanzania-Tourism-Prediction

Tanzania Tourism Prediction Challenge # 3-Tanzania-Tourism-Prediction # Module 3 Zindi Challenge – Tanzania Tourism Prediction **Due:** Monday by 9:59 **Points:** 100 **Submission:** File upload **Availability:** 16 Feb at 10:00 – 24 Feb at 9:59 --- ## 📘 Challenge Overview This project is part of **Module 3** in the Zindi Challenge series. The objective is to apply **data preprocessing** and **feature engineering** skills to the **Tanzania Tourism Prediction Challenge**. By the end, you should have a clean dataset with useful features ready for modeling. --- ## 📝 Instructions ### Step 1: Join and Download the Data - Join the Tanzania Tourism Prediction Challenge as a team (pair partner). - Download the datasets: - `Train.csv` - `Test.csv` - `SampleSubmission.csv` - Place them in your project folder. ### Step 2: Explore the Data - Load the training dataset into a Jupyter Notebook. - Use **Pandas** to inspect the dataset: - Check shape, preview rows, and review data types. - Summarize with `.describe()`, `.info()`, and `.value_counts()` for categorical variables. ### Step 3: Handle Missing Data and Outliers - Identify missing values using `.isnull().sum()`. - Apply strategies to handle them (drop, mean, median, or mode). - Detect outliers (e.g., using interquartile range). Decide whether to keep, remove, or transform them. ### Step 4: Encode Categorical Variables - Apply **label encoding** to at least one binary column. - Apply **one-hot encoding** to at least one multi-category column. - Compare the dataset before and after encoding. ### Step 5: Create New Features - Engineer at least **two new features** from existing data. Examples: - Combine multiple columns into a single metric. - Extract useful information (e.g., age groups, ratios). - Explain why your new features could improve predictions. ### Step 6: Reflection Write a short (1–2 pages) reflection note as a team. Include: - Key challenges in cleaning, encoding, and …