Toyota Ultimate Racing Division (TURD)

Inspiration

In today’s digital economy, online reviews shape consumer decisions and business reputations. We noticed that many location-based reviews are low-quality, irrelevant, or misleading, making it hard for users and businesses to trust them. Our inspiration was to create a tool that filters out the noise and highlights reviews that truly matter.

What it does

Our project scrapes reviews from Google Maps, cleans and processes the text, analyzes sentiment, and classifies reviews by quality and relevance. Users can explore trends, keyword clouds, and sentiment insights, helping them make better-informed decisions.

How we built it

We used the Google Maps API to gather reviews and combined NLP techniques with machine learning for analysis. Python libraries like pandas, nltk, scikit-learn, and plotly powered our data handling, sentiment analysis, classification, and visualizations. Interactive widgets allowed users to explore the data easily.

Challenges we ran into

Working with noisy, unstructured review data was tricky. Filtering out spammy content while retaining meaningful reviews required careful preprocessing and feature engineering. Fine-tuning our machine learning models for reliable relevance classification was also a challenge.

Accomplishments that we're proud of

We built a pipeline that automatically scrapes and analyzes reviews, identifies high-quality content, and presents interactive visualizations. Our supervised models can distinguish relevant reviews with impressive accuracy, making insights accessible and actionable.

What we learned

We gained hands-on experience in NLP, sentiment analysis, and machine learning workflows. We also learned the importance of clean, well-labeled datasets and the value of interactive visualizations in communicating insights effectively.

What's next for Toyota Ultimate Racing Division (TURD)

Next, we plan to expand our dataset, improve model accuracy, and integrate real-time review analysis. We also aim to enhance the interactive interface so users can filter reviews by relevance, sentiment, and location with even greater ease.