Sentiment analysis of Deepnote socials 💻

In this data app, Deepnote's market trends on social media will be explored by analyzing discussions and mentions across platforms like X (formerly Twitter), Hacker News, Reddit, Stackoverflow, GitHub, and LinkedIn, and their associated sentiment.

topic

data_source

Twitter

In the graph below, a filter was set for the period between April and June, where the number of posts about Deepnote can be seen with a granularity of one week.

In the code snippet bellow the NTKL library is used - it provides tools for processing text.

#NLTK library - is a Python library used for working with human language data nltk.download('stopwords') def preprocess_text(text): # Remove HTML text = re.sub(r'<.*?>', '', text) # Remove URLs text = re.sub(r'http\S+', '', text) # Remove special characters text = re.sub(r'[^A-Za-z\s]', '', text) # Convert to lowercase letters text = text.lower() # Remove stopwords stop_words = set(stopwords.words('english')) text = ' '.join([word for word in text.split() if word not in stop_words]) return text text_field = 'summary' df_1['clean_text'] = df_1[text_field].apply(preprocess_text)

Sentiments

Positive (close to 1), neutral (around zero), and negative (close to -1).

Good sentiment (0.3644)

I believe Deepnote is best in class here, I know you said you've tried it, but there's really nothing better on the market. Maybe hyperquery? But they are shutting down.

Neutral sentiment (0)

Bad sentiment (-0.7883)

week intro deepnote changes

deepnotehq hey keep getting problem loading projects list projects failed go loaded error cant access projects please help fix

Visualization

In the graphs below, the sentiment trends in the dataset spanning from March to July 2024 are analyzed. The first graph illustrates the average sentiment score over time, plotted monthly. This line graph provides an overview of how sentiment fluctuates throughout the specified period. The second visualization, a histogram, shows the distribution of sentiment scores across the dataset. In the third pie chart, the percentage representation of positive, neutral, and negative sentiments is shown.

Calculation of the percentage representation of positive, neutral, and negative sentiment:

In the graph below, the average sentiment for the social networks Reddit, Twitter, Stack Overflow and HackerNews can be seen - covering the period from March to July 2024.

# Check if each entry in 'clean_text' has at least 10 words df_1['word_count'] = df_1['clean_text'].apply(lambda x: len(x.split())) # Filter the DataFrame to only include entries with at least 10 words df_filtered = df_1[df_1['word_count'] >= 10] if not df_filtered.empty: all_text = ' '.join(df_filtered['clean_text']) # Remove the word "OH" stopwords = set(STOPWORDS) stopwords.add("OH") # Set the color to use only shades of blue def blue_color_func(word, font_size, position, orientation, random_state=None, **kwargs): return "hsl({}, 70%, 50%)".format(np.random.randint(200, 250)) # Shades of blue meta_mask = np.array(Image.open('./deepnote_logo.png')) # Word cloud generation wordcloud = WordCloud( width=200, height=200, # background_color='white', mask=meta_mask, color_func=blue_color_func, stopwords=stopwords, max_words=200, min_font_size=12, max_font_size=100, random_state=42 ).generate(all_text) # Display the word cloud plt.figure(figsize=(10, 5)) plt.imshow(wordcloud, interpolation='bilinear') plt.axis('off') plt.show() else: print("Not enough words to generate a word cloud.")