Social Media Sentiment Analysis Data Science Project

"""Transparent lexicon-based sentiment analysis for synthetic social posts."""

from __future__ import annotations

import re
import pandas as pd

POSITIVE = {"helpful", "love", "excellent", "fast", "clear", "great", "easy", "happy"}
NEGATIVE = {"slow", "confusing", "broken", "bad", "late", "difficult", "unhappy", "error"}


def demo_posts() -> pd.DataFrame:
    return pd.DataFrame({
        "post": ["The tutorial was clear and helpful", "Upload is slow and shows an error", "Great support and fast reply", "The layout is confusing", "It works", "I love the easy examples"],
        "likes": [18, 7, 23, 5, 3, 31],
    })


def score_text(text: str) -> tuple[int, str]:
    tokens = set(re.findall(r"[a-z]+", text.lower()))
    score = len(tokens & POSITIVE) - len(tokens & NEGATIVE)
    label = "positive" if score > 0 else "negative" if score < 0 else "neutral"
    return score, label


def analyse_posts(data: pd.DataFrame) -> tuple[pd.DataFrame, pd.DataFrame]:
    result = data.copy()
    scored = result["post"].map(score_text)
    result[["sentiment_score", "sentiment"]] = pd.DataFrame(scored.tolist(), index=result.index)
    summary = result.groupby("sentiment", as_index=False).agg(posts=("post", "size"), average_likes=("likes", "mean"))
    return result, summary


def main() -> None:
    results, summary = analyse_posts(demo_posts())
    print(results.to_string(index=False))
    print("\nSummary:\n", summary.round(2).to_string(index=False))
    print("Rule-based labels miss sarcasm, context, dialect, and mixed sentiment.")


if __name__ == "__main__":
    main()