📊 Dropbox User Sentiment Analysis using Airbyte 🪼 and Motherduck 🦆

In this blog, we'll dive 🌊🤿 into building a Streamlit-based dashboard for analyzing Dropbox User Sentiment using Airbyte for data extraction and Motherduck (DuckDB) for storage and querying. This post continues from our previous discussion in "Leveraging Airbyte 🪼 and Motherduck 🦆 for Sentiment Analysis" and explores how these technologies integrate with Streamlit to create an interactive and insightful data analysis application.

---

📁 Folder Structure Overview

DROPBOX-REVIEWS-ANALYSIS
├── .devcontainer
│   ├── devcontainer.json
├── .streamlit
│   ├── config.toml
├── assets
│   ├── main.png
├── dropbox-reviews-analytics
│   ├── src
│   │   ├── config
│   │   │   ├── __init__.py
│   │   │   ├── config.py
│   │   ├── utils
│   │   │   ├── __init__.py
│   │   │   ├── database.py
│   │   ├── app.py
├── .env
├── venv
├── .gitignore
├── LICENSE.md
├── README.md
├── requirements.txt

• .devcontainer/devcontainer.json: Configures development environment.
• .streamlit/config.toml: Streamlit's UI style and configuration.
• assets: Stores static assets like images.
• src/config/config.py: Handles environment variables.
• src/utils/database.py: Queries data from Motherduck.
• src/app.py: Streamlit dashboard and logic.
• .env: Stores environment variables securely.

👉 Tips On Folder Structure

• DROPBOX-REVIEWS-ANALYSIS: This is the outer folder of repo on Github (while building project on your own don't create this folder)
• dropbox-reviews-analytics: This is the main folder where you will add src followed by config and utils

---

🎏 Streamlit Setup

Streamlit is an open-source Python library that enables developers to create interactive web apps for data science and machine learning projects.

📜 Code Snippet: Streamlit Core Structure

import streamlit as st
import plotly.express as px
from utils.database import get_reviews_for_sentiment

st.set_page_config(page_title="Dropbox Analysis", page_icon="🗳️", layout="wide")

# Title
st.markdown("## 🗳️ Dropbox Sentiment Analysis")

# Sidebar
sentiment_type = st.sidebar.selectbox("Sentiment Analysis Type", ["Polarity", "Subjectivity"])

# Fetch and Display Data
reviews_df = get_reviews_for_sentiment()
st.dataframe(reviews_df)

When you run app.py with streamlit run src/app.py, the dashboard launches at http://localhost:8501.

---

📊 Core Logic of Sentiment Analysis

Sentiment analysis is powered by TextBlob to determine the polarity (positive/negative sentiment) and subjectivity (factual/opinionated content) of reviews.

🧠 Sentiment Analysis Function

from textblob import TextBlob

def get_sentiment(text):
    blob = TextBlob(str(text))
    return blob.sentiment.polarity if sentiment_type == "Polarity" else blob.sentiment.subjectivity

📈 Visualization Example

fig = px.histogram(reviews_df, x='sentiment', title='Sentiment Distribution')
st.plotly_chart(fig)

---

🦆 *Database Integration with Motherduck *

🔗 database.py

import duckdb
from config.config import MOTHERDUCK_TOKEN

def get_connection():
    return duckdb.connect(f"md:?token={MOTHERDUCK_TOKEN}")

def get_reviews_for_sentiment():
    conn = get_connection()
    query = """
    SELECT content, score FROM dropbox_reviews WHERE content IS NOT NULL
    """
    return conn.execute(query).fetch_df()

This code fetches Dropbox review data securely using MOTHERDUCK_TOKEN stored in .env through config.py file.

🗂️ config.py

import os
from dotenv import load_dotenv

load_dotenv()
MOTHERDUCK_TOKEN = os.getenv("MOTHERDUCK_TOKEN")

🔄 Connection Between app.py and database.py

The app.py imports get_reviews_for_sentiment from database.py, creating a seamless flow of data into the dashboard.

---

⚙️ Why devcontainer.json and config.toml?

• devcontainer.json: Provides a consistent environment for development, anyone willing to use Docker for containerization can use it.
• config.toml: Controls Streamlit UI customization (e.g., colors, fonts, themes).

Example config.toml:

[theme]
primaryColor="#0061FE"
backgroundColor="#0E1117"
secondaryBackgroundColor="#262730"
textColor="#FAFAFA"
font="Monospace"

---

⚠️ Deployment Challenges

• Avoid specifying exact library versions in requirements.txt, like instead of plotly == 5.24.1 write plotly only.
• Ensure .env is configured correctly in deployment environments.
• Validate database connection tokens during runtime.

Backend Deployment Flow:

1. Load environment variables from .env.
2. Establish connection with Motherduck DB.
3. Fetch and process data.
4. Render dashboard in Streamlit.

---

🎯 Conclusion

We successfully built a Dropbox Reviews Sentiment Analysis Dashboard using Airbyte, Motherduck, and Streamlit. This project demonstrates the power of data analysis and visualization.

👨‍💻 Check out the complete code on GitHub.
📺 Live PROJECT https://airbyte-motherduck-hackathon-sentiment-analysis.streamlit.app

Sentiment Analysis #Happy Coding! #Airbyte🪼🦆