Data Science Portfolio: Tweet Analysis

This repository contains two Python scripts demonstrating data science skills in tweet analysis. These projects were originally developed as part of Codecademy's portfolio-building exercises.

Scripts Overview

1. viral_tweets.py

• Objective: Predict whether a tweet will go viral based on its features such as length, follower count, and retweet count.
• Techniques Used:
  • Feature engineering
  • Data preprocessing (scaling)
  • K-Nearest Neighbors (KNN) classifier to determine optimal k value.
• Libraries: pandas, numpy, scikit-learn, matplotlib

2. tweet_location.py

• Objective: Classify tweets' geographic origin (New York, London, or Paris) based on their text content.
• Techniques Used:
  • Text preprocessing and vectorization using CountVectorizer
  • Naive Bayes classification for text data.
• Libraries: pandas, scikit-learn

Project Highlights

• These scripts showcase practical applications of data preprocessing, feature engineering, and machine learning algorithms.
• The viral_tweets.py script demonstrates predictive modeling and hyperparameter tuning using KNN.
• The tweet_location.py script highlights text-based classification with Naive Bayes.

How to Use

1. Clone the repository.
2. Ensure you have Python 3.6+ installed.
3. Install the required libraries using pip install -r requirements.txt.
4. Place the necessary JSON datasets (random_tweets.json, new_york.json, london.json, paris.json) in the same directory as the scripts. These datasets are included in the zip file provided.
5. Run the scripts:
   • python viral_tweets.py
   • python tweet_location.py

About

These projects are part of Codecademy's data science portfolio, completed to build foundational skills in data analysis and machine learning.

---

Feel free to explore the scripts and reach out with any questions or suggestions!

Author: Rayan Roshan