📦 Updates

Author: CSFelix

4 - Python/4 - Data Science/1 - Statistic Analysis/0 - codes/10 - DataPrep Library.py

@@ -0,0 +1,17 @@
+# pip install dataprep
+from dataprep.eda import create_report
+from dataprep.eda import plot
+from dataprep.eda import plot_correlation
+from dataprep.eda import plot_missing
+
+# Full Report
+create_report(df)
+
+# Plots
+plot(df)
+
+# Correlations
+plot_correlation(df)
+
+# Missing Values
+plot_missing(df)

4 - Python/4 - Data Science/1 - Statistic Analysis/0 - codes/11 - WordCloud.py

@@ -0,0 +1,21 @@
+# Libraries
+import pandas as pd
+import matplotlib.pyplot as plt
+from wordcloud import WordCloud # pip install wordcloud
+import numpy as np
+
+# Function to generate word clouds
+def generate_wordcloud(df, column, max_font_size=60, mask, collocations=False):
+		text = df[column].values
+		return WordCloud(
+			max_font_size=max_font_size
+			, collocations=collocations).generate(str(text)
+			, mask=mask
+		)
+
+# Generating and Plotting Word Cloud
+mask = np.array(Image.open(path.join(d, "stormtrooper_mask.png")))
+wordcloud = generate_wordcloud(df, 'video_title') # or any string feature
+plt.imshow(wordcloud)
+plt.axis('off')
+plt.show()