Customer Review Sentiment Analysis NLP
Budget: ₹750 – ₹1,250 INR
I need an AI workflow that reads batches of customer reviews and returns reliable sentiment scores I can plug straight into my reports. The core of the project is natural language processing, specifically sentiment analysis, so experience with popular frameworks such as Python + NLTK, spaCy, Hugging Face Transformers, or similar will be essential.
You’ll start from raw text (CSV or JSON, whichever is easier for you) and deliver a clean, documented pipeline that:
• cleans and tokenises the reviews,
• trains or fine-tunes a sentiment model (I’m open to pre-trained BERT, RoBERTa, etc.),
• outputs a simple table with the original review, predicted polarity (positive / neutral / negative) and a confidence score.
I’ll measure success by the model’s accuracy on a held-out validation set and by how easily I can rerun the code on new review data. A short read-me and example notebook will round everything off so I can reproduce results on my side.
You’ll start from raw text (CSV or JSON, whichever is easier for you) and deliver a clean, documented pipeline that:
• cleans and tokenises the reviews,
• trains or fine-tunes a sentiment model (I’m open to pre-trained BERT, RoBERTa, etc.),
• outputs a simple table with the original review, predicted polarity (positive / neutral / negative) and a confidence score.
I’ll measure success by the model’s accuracy on a held-out validation set and by how easily I can rerun the code on new review data. A short read-me and example notebook will round everything off so I can reproduce results on my side.