Transform your business with Analysr
This is my submission for Airbyte-Motherduck Hackathon - December 2024 - January 2025
Here’s a speedy summary:
- 1.0.0
- With your customer reviews in Motherduck, along with your chosen business stack and areas of interest, Analysr is ready to dish out some insightful analytics. To sweeten the deal, Groq is also integrated to help you navigate all your growth phases.
- Your analytics lineup features Aspect Analysis, a Word Sentiment Heatmap (for those feelings), Advanced Text Analysis, Groq Business Analytics, Keyphrase Analysis, and a handy Competitive Advantage.
- Check it out at:
https://growwithanalysr.web.app/ - Productionhttps://growwithanalysr-staging.vercel.app/ - Experimental, for new features
- Scale Beyond Regular AI Capabilities: Traditional AI systems, like ChatGPT, struggle to handle extensive datasets (e.g., 65,000+ records) effectively. Analysr bridges this gap.
- Seamless Motherduck, Airbyte, Groq Integration: Thanks to Motherduck wasm client, Airbyte's API and Groq SDK.
- Data-Driven Insights: By combining AI with visualization tools, Analysr allows users to uncover trends, anomalies, and actionable insights quickly and intuitively.
- User-Friendly Visualization: Visual AI integration transforms raw data into understandable and compelling graphics, enabling better decision-making.
- Streamlined Process: Reduces reliance on multiple tools by offering an all-in-one platform for schema analysis and visualization.
- To obtain customer review insights, sync your data to Motherduck with the schema:
{ "review_text": "string", "stars": "number" }(More schemas support are in the future roadmap). We recommend using Airbyte due to its extensive list of sources and seamless data movement.
- Visit the Analysr website at (growwithanalysr.web.app) and click on the "Get Started Now" button for onboarding.
- Select your business stack and substack; Groq and queries will use this information to fetch insights.
- Input your Motherduck token and wait for the connection to be established (the time required will depend on your network bandwidth).
- Select the database and table where your customer reviews or any related reviews exist, and set the data limit.
- Input your Groq token (recommended) to obtain AI-based insights.
- Optionally provide your Airbyte bearer token (from the cloud.airbyte.com settings page) and connection ID (from the connections tab URL) to trigger a sync for updating your Motherduck table.
- Finally, input your area of interest for insights, such as customer satisfaction, and click "Continue to Dashboard."
- Wait a few seconds until all queries are executed and visualized.
- Voilà! Your dashboard will be ready, featuring all of Analysr's capabilities to support your next big step!
Need a dataset and one example method to test?
- Hugging face dataset URL which I used - https://huggingface.co/datasets/Yelp/yelp_review_full
- Import it to Motherduck via Airbyte (Set huggingface as source and Motherduck as destination) OR attach using my share link
-- Run this snippet to attach the database
ATTACH 'md:_share/my_db/de60469b-3a05-4d74-bf63-4c1549dd55b6';- Get a Groq token at, https://console.groq.com/keys
- Click on Continue to the dashboard! That's it. Please try it yourself, it's fun!
- Aspect Analysis: Gain insights into different aspects of customer feedback.
- Word Sentiment Heatmap: Visualize sentiment trends in your reviews.
- Advanced Text Analysis: Delve deeper into the nuances of customer language.
- Groq Business Analytics: Access data-driven insights to inform your growth strategy.
- Keyphrase Analysis: Identify and analyze key phrases that matter to your customers.
- Competitive advantage: Benchmark your reviews to know your current positive/negative sentiment advantage.
- Frontend: React, TypeScript, Tailwind CSS, Vite
- Analytics: MotherDuck (DuckDB), GROQ AI
- Data Integration: Airbyte
- Visualization: Recharts
- State Management: Zustand
- Animations: Framer Motion
- Hosting: Firebase (Production), Vercel (Experiment)
- Proxy: Supabase edge functions
- CI/CD: GitHub Actions for automated deployment
Declarations: For development, the VSCode code editor, Codeium AI helper extension, and suggestions from ChatGPT were used.
- Microservice for generating queries: Currently all queries for analytics are highly coupled with code, separation of concerns and transition to microservice is required
- Create express server proxy and deploy as superbase functions
- Separate DuckDB queries as an API call response
- Enhance microservice with GPT Wrapper
- Improve business insights from Groq: At present, it produces some inaccuracies due to the limitations of the open-source
mixtralmodel, which lacks the necessary funding to enhance its capabilities.
- Clone the repository:
git clone https://github.com/btkcodedev/airbyte-motherduck-hackathon.git- Install dependencies:
npm install- Start the development server:
npm run dev- Secure token management
- Row-level security
- Data encryption
- Secure API endpoints
MIT License - feel free to use this project for your purposes!
Contributions are welcome! Please feel free to submit a Pull Request.
For questions and support, please open an issue or contact at [email protected]
Built with ❤️ for growing businesses by
btkcodedev