Add files via upload

Author: idilersudas

Crypto_Sentiment_Analysis_VADER.xlsx

@@ -0,0 +1,208 @@
+import pandas as pd
+from textblob import TextBlob
+
+notebook_path = "/Users/umuteyidogan/Desktop/IGP_Project/crypto_headline_data.csv"
+
+df = pd.read_csv(notebook_path)
+
+crypto_lexicon = {
+    'bullish': 2.0,
+    'bearish': -2.0,
+    'hodl': 1.5,
+    'fomo': -1.5,
+    'pump': 1.5,
+    'dump': -1.5,
+    'moon': 2.0,
+    'whale': 0.5,
+    'altcoin': 0.5,
+    'scam': -2.5,
+    'rugpull': -2.0,
+    'pump and dump': -2.0,
+    'moonshot': 2.0,
+    'to the moon': 2.0,
+    'bear market': -2.0,
+    'bull market': 2.0,
+    'crypto': 0.5,
+    'blockchain': 0.5,
+    'yield farming': 1.0,
+    'staking': 1.0,
+    'token': 0.5,
+    'gas fee': -0.5,
+    'mining': 0.5,
+    'hashrate': 0.5,
+    'volatile': -1.5,
+    'regulation': -0.5,
+    'adoption': 1.5,
+    'innovation': 2.0,
+    'security': 1.0,
+    'fraud': -2.5,
+    'hack': -2.0,
+    'partnership': 1.5,
+    'investment': 1.5,
+    'exchange': 0.5,
+    'wallet': 0.5,
+    'halving': 1.5,
+    'funding': 1.0,
+    'launch': 1.5,
+    'collapse': -2.5,
+    'lawsuit': -2.0,
+    'profit': 2.0,
+    'loss': -2.0,
+    'growth': 1.5,
+    'decline': -1.5,
+    'risk': -1.0,
+    'opportunity': 2.0,
+    'recovery': 1.5,
+    'crash': -2.5,
+    'surge': 2.0,
+    'plummet': -2.0,
+    'rebound': 1.5,
+    'stable': 1.0,
+    'plunge': -2.0,
+    'airdrop': 1.0,
+    'bull': 1.5,
+    'bear': -1.5,
+    'fud': -1.5,
+    'rekt': -2.0,
+    'satoshi': 0.5,
+    'burn': 1.0,
+    'mint': 1.0,
+    'whitelist': 1.0,
+    'blacklist': -1.0,
+    'whale': 0.5,
+    'staking reward': 1.0,
+    'buy the dip': 2.0,
+    'sell the news': -1.0,
+    'short squeeze': 1.5,
+    'margin call': -1.5,
+    'paper hands': -1.5,
+    'diamond hands': 1.5,
+    'moonboy': 1.5,
+    'bagholder': -1.5,
+    'bear trap': -1.5,
+    'bull trap': -1.5,
+    'dead cat bounce': -1.5,
+    'double top': -1.5,
+    'double bottom': 1.5,
+    'cup and handle': 1.5,
+    'head and shoulders': -1.5,
+    'golden cross': 1.5,
+    'death cross': -1.5,
+    'consolidation': 0.5,
+    'take profit': 1.5,
+    'breakout': 1.5,
+    'breakdown': -1.5,
+    'bull flag': 1.5,
+    'bear flag': -1.5,
+    'buy wall': 1.5,
+    'sell wall': -1.5,
+    'stop-loss': 0.5,
+    'take profit': 1.5,
+    'recession': -1.0,
+    'economic downturn': -2.0,
+    'market correction': -1.5,
+    'bull run': 2.0,
+    'bear run': -2.0,
+    'alt season': 1.5,
+    'defi': 0.5,
+    'dao': 1.0,
+    'dapp': 1.0,
+    'nft': 0.5,
+    'smart contract': 0.5,
+    'ico': 1.5,
+    'fud': -2.0,
+    'ath': 2.0,
+    'atl': -2.0,
+    'btc': 0.5,
+    'xbt': 0.5,
+    'confirmation': 0.5,
+    'cold storage': 1.0,
+    'consensus': 0.5,
+    'cross-chain': 0.5,
+    'cryptography': 0.5,
+    'decryption': 0.5,
+    'dominance': 0.5,
+    'double spending': -2.0,
+    'dusting attack': -2.0,
+    'emission': 0.5,
+    'encryption': 0.5,
+    'impermanent loss': -1.5,
+    'memecoin': -1.0,
+    'node': 0.5,
+    'order book': 0.5,
+    'parachains': 0.5,
+    'peer-to-peer (P2P)': 0.5,
+    'phishing attack': -2.0,
+    'smart contract': 0.5
+
+
+}
+
+def get_custom_sentiment(text, lexicon):
+    if isinstance(text, str):
+        words = text.split()
+        sentiment_score = 0.0
+        for word in words:
+            if word in lexicon:
+                sentiment_score += lexicon[word]
+            else:
+                analysis = TextBlob(word).sentiment
+                sentiment_score += analysis.polarity
+        return sentiment_score / len(words) if words else 0
+    return 0
+
+df['Sentiment'] = df['Headline'].apply(lambda x: get_custom_sentiment(x, crypto_lexicon))
+
+# Assign sentiment labels
+def assign_label(score):
+    if score > 0:
+        return 1
+    elif score < 0:
+        return -1
+    else:
+        return 0
+
+df['Sentiment_Label'] = df['Sentiment'].apply(assign_label)
+
+print(df)
+
+ #Aggregating daily sentiment
+df['Published date'] = pd.to_datetime(df['Published date']).dt.date  # Ensure the date column is in datetime format
+daily_sentiment = df.groupby('Published date').agg(
+    Positive_Count=('Sentiment_Label', lambda x: (x == 1).sum()),
+    Negative_Count=('Sentiment_Label', lambda x: (x == -1).sum()),
+    Neutral_Count=('Sentiment_Label', lambda x: (x == 0).sum())
+).reset_index()
+
+# Calculate percentages
+daily_sentiment['Total'] = daily_sentiment['Positive_Count'] + daily_sentiment['Negative_Count'] + daily_sentiment['Neutral_Count']
+daily_sentiment['Positive_Percentage'] = (daily_sentiment['Positive_Count'] / daily_sentiment['Total']) * 100
+daily_sentiment['Negative_Percentage'] = (daily_sentiment['Negative_Count'] / daily_sentiment['Total']) * 100
+daily_sentiment['Neutral_Percentage'] = (daily_sentiment['Neutral_Count'] / daily_sentiment['Total']) * 100
+
+print(daily_sentiment)
+
+# Save the results
+
+output_path = "/Users/umuteyidogan/Desktop/IGP_Project/Crypto_Sentiment_Analysis.xlsx"
+df.to_excel(output_path, index=False)
+
+print(f"Analysis results have been saved to {output_path}")
+
+# Save daily sentiment to a new file
+daily_sentiment_output_path = "/Users/umuteyidogan/Desktop/IGP_Project/Daily_Sentiment_Analysis.xlsx"
+daily_sentiment.to_excel(daily_sentiment_output_path, index=False)
+print(f"Daily sentiment resultsHere's the updated code that includes the columns for daily sentiment aggregation and calculates the percentages of positive, negative, and neutral sentiments:")
+
+'''
+import matplotlib.pyplot as plt
+
+plt.figure(figsize=(10, 6))
+plt.hist(df['Sentiment'], bins=20, edgecolor='k', alpha=0.7)
+plt.title('Sentiment Analysis Distribution')
+plt.xlabel('Sentiment Polarity')
+plt.ylabel('Frequency')
+plt.grid(True)
+plt.show()  
+
+'''