Data Engineering: Loaded the Kaggle social-media sentiment data (Twitter, Facebook, Instagram) and ensured no missing values. Added a Clean_Text column by lower-casing, removing punctuation and common words, and stemming, so the dataset is ready for machine-learning.

Predictive Modeling: Tested four classifiers and selected SVM (linear kernel) as the champion, delivering 71 % overall accuracy and ~90 % precision on negative sentiment, outperforming Naive Bayes, Random Forest, and Logistic Regression.

Audience & Platform Insights: Discovered that the USA, Canada, and UK generate triple the positive chatter of other nations, while Instagram (used by 36 % of these users) hosts the highest positivity, flagging it as the prime venue for upbeat brand engagement.

Timing Playbook: Uncovered clear temporal swings ; June tops positive sentiment, September peaks in negativity, and Sundays outperform Saturdays for optimism, informing when to launch campaigns or deploy supportive messaging.