Unlocking Insights: Analyzing Google Reviews with LLMs

Terry Tan -

Customer reviews have become an integral part of the decision-making process for consumers. When it comes to understanding the reputation and quality of a business, Google Reviews serve as a valuable source of information. However, analyzing and extracting meaningful insights from a large volume of reviews can be a time-consuming and daunting task.

SerpApi has developed an Open Source project that leverages the power of OpenAI LLM and Google APIs as the real time data source - Google Reviews Analyzer, which aim to make the process more efficient.

For business owner looking to understand customer sentiment and feedback, this tool can save you hours of manual review analysis and provide you with actionable information. This can be useful to consumers as well, as it provides them with valuable insights and aids in making informed decisions based on the experiences of others.

How It Works

The key component to intelligence is to have a vast amount of data. In the context of intelligent systems such as the Google Reviews Analyzer, having access to a substantial dataset of Google Reviews enables the algorithm to capture the nuances of customer sentiment, identify trends, and extract meaningful insights. This tool employs SerpApi's Google Maps Reviews API and using pagination it can extract a vast amount of reviews data from a business, it is very important for making accurate analysis.

OpenAI LLM (Language Models) indeed serves as the main component that makes the Google Reviews Analyzer project possible. The advanced capabilities of OpenAI LLM play a crucial role in analyzing and summarizing the Google Reviews data effectively. With its natural language processing capabilities, OpenAI LLM enables the system to understand and interpret the content of the reviews, extract key information, and identify sentiment and patterns.

Sample prompt:

Reviews:
- I visited here because apparently it's an iconic establishment in Portland. I'm not much of a coffee drinker, unless I'm absolutely exhausted, but it was chill just crossing it off my list. I don't know if it is worth the drive from out of town tho. Maybe make it a part of your day if you're exploring Portland instead because it definitely is hard to find parking. I parked near a hydrant, thank goodness I didn't get a ticket, just for this opportunity. I purchased the decaf ice latte based upon the recommendation of the worker there. Chill atmosphere! Vegetarian options: Oat milk. Parking: Hard to find parking. I recommend traveling here on foot. Dietary restrictions: Oat milk is the default. But I'm pretty sure they can accommodate to a point.
- Nice ambience, friendly staff and good coffee. I liked the setup of this place. Very clean and well maintained as well. We ordered Caramel Latte, Matcha latte, banana bread, plain croissant and pista croissant! It’s pet friendly and has a good lobby seating but very limited patio seating.
- ...many more reviews
    
---
Create 10 most common labels for these reviews and give rating of 1 to 5 ⭐, with 1 ⭐ being the lowest rating and 5 ⭐ being the highest rating, it should be based on the occurrence rate, and if the label is negative as 👎 or positive as 👍.

Example:
Label Name (positivity): ⭐ emoji
Sample prompt that is used

How To Use

Before running the application, we have to get the necessary API key:

  1. Duplicate the file .env.example and rename it to .env
  2. We need 2 API keys from SerpApi and OpenAI for SERPAPI_KEY and OPENAI_API_KEY respectively. You can sign up for SerpApi and get 100 free credits.

Installation

First, install the dependencies

npm install

Run the development server:

npm run dev

Open http://localhost:3000 with your browser to see the result.

Future Improvements

We have another well made project similar to this - SerpApi's ChatGPT Review Analyzer Chrome Extension which is a Chrome Extension that utilizes ChatGPT to access the OpenAI's LLM.

Building this as a Web App I think there is more we can do. We can make enable chat based experience to interact with the data that allow business owners to get more insights, like having the experience to interview the customers. I think there are more possibilities than what I have right now.

Conclusion

The power of LLMs unlocks new possibilities that were not easily attainable before. With it, Google Reviews Analyzer project can delve deeper into the nuances of customer sentiment, extract meaningful insights, and summarize vast amounts of reviews in a fraction of the time it would take humans to do so manually or huge resources to make proprietary software. The ability of LLMs to comprehend context, identify patterns, and infer meaning allows the project to uncover valuable information that may have been previously overlooked or time-prohibitive to analyze.

Feel free to reach out to me at my email if you have any interesting ideas or any questions.