Sentiment analysis
Sentiment analysis is the computational identification, classification, or measurement of opinions, attitudes, emotions, or evaluative polarity expressed in text or other data.
Starting Data Analytics with Generative AI and Python
Sentiment analysis is the computational analysis of text to determine expressed opinions, emotions, or evaluative polarity.
- How it is used
- Classifies customer feedback to identify satisfaction, dissatisfaction, and issues worth investigating.
- How deep it goes
- Works through keyword, lexicon, API, and pretrained-model approaches, then compares sensitivity and specificity without detailing model internals.
- Examples
- Uses Portuguese product reviews, a mixed review, Python API calls, and a 500-review method comparison.
- What you can do after
- Apply it via Python and an API or pretrained classifier, then evaluate labels against review scores and inspect questionable cases.
Read how Starting Data Analytics with Generative AI and Python explains sentiment analysis →
Natural Language Processing in Action, Second Edition
Sentiment analysis is the computational determination of opinions, emotions, or evaluative attitudes expressed in text.
- How it is used
- Deploys sentiment analysis to extract emotional signals from customer feedback, filter harmful messages, and support conversational systems.
- How deep it goes
- Explains sentiment as token information and numerical output, while stopping short of detailed model or algorithm implementation.
- Examples
- Uses product reviews, open-ended customer comments, star ratings, junk mail, and troll messages as concrete applications.
- What you can do after
- Apply it via an NLP pipeline to convert large collections of feedback into emotional ratings and support message filtering.
Read how Natural Language Processing in Action, Second Edition explains sentiment analysis →
Data Without Labels
Sentiment classification is the assignment of text to categories representing expressed opinions, attitudes, or emotional polarity.
- How it is used
- Gauges customer satisfaction and pain points in review text when written opinions may diverge from star ratings.
- How deep it goes
- Walks through review collection, cleaning, polarity scoring, threshold labels, and plots; uses TextBlob rather than deriving or training a classifier.
- Examples
- Uses Amazon thermostat reviews, star-rating counts, TextBlob scores, satisfaction bands, and a ratings-versus-sentiment figure.
- What you can do after
- Apply it via TextBlob to score cleaned reviews, assign satisfaction labels, and compare their distribution with star ratings.
Grokking Machine Learning
The computational task of determining an expressed opinion or affective polarity in text.
- How it is used
- Frames sentiment analysis as classifying text into positive or negative categories.
- How deep it goes
- Demonstrates word-count features, logistic-classifier training, and inspection of learned coefficients.
- Examples
- Uses labeled IMDB movie reviews and learned coefficients for words such as “comfortable.”
- What you can do after
- Apply it via Turi Create to turn review text into word-count features and train a logistic classifier.
Read how Grokking Machine Learning explains sentiment analysis →
The Complete Obsolete Guide to Generative AI
Sentiment analysis is the computational classification of text according to expressed affective polarity such as positive, negative, or neutral.
- How it is used
- Measures large-scale opinion in reviews and social posts to surface consumer, public, and market sentiment.
- How deep it goes
- Explains polarity classification and preprocessing, then tests prompted GPT labels against pre-labeled posts without building a classifier from scratch.
- Examples
- Uses 1,000 pre-labeled X product-review posts, Python CSV processing, GPT-3/GPT-3.5 prompts, and prediction-frequency counts.
- What you can do after
- Apply it via GPT prompts to label consumer comments, compare outputs with known labels, and manually verify unreliable results.
Read how The Complete Obsolete Guide to Generative AI explains sentiment analysis →
First-Party Data Activation
The computational or analytical assessment of expressed attitudes and emotions in customer communications.
- How it is used
- Uses customer communications to support proactive service, product improvement, marketing insight, and message testing.
- How deep it goes
- Classifies sentiment as positive, negative, or neutral, then extends toward generative summaries and deeper insights without implementation detail.
- Examples
- Uses email, social media, reviews, support tickets, marketing messages, and artificial customer personas as analysis inputs.
- What you can do after
- Apply it via sentiment analysis tools to monitor satisfaction, prioritize concerns, improve products, and test marketing responses.
Read how First-Party Data Activation explains sentiment analysis →
Microsoft Azure in Action
Sentiment analysis is the computational determination of opinions, emotions, or attitudes expressed in text.
- How it is used
- Uses sentiment analysis to understand customer opinions in feedback and inform improvements to a bookseller’s offerings.
- How deep it goes
- Introduces the purpose through a customer-feedback scenario but provides no sentiment-analysis procedure or output.
- Examples
- Uses a bookseller’s customer feedback as the scenario for identifying opinions.
- What you can do after
- Evaluate/understand it as a way to learn what customers think before improving offerings.
Read how Microsoft Azure in Action explains sentiment analysis →
Snowflake Data Engineering
Sentiment classification assigns text to categorical sentiment labels such as negative, neutral, or positive.
- How it is used
- Classifies customer reviews so teams can prioritize specific feedback and monitor sentiment distribution over time.
- How deep it goes
- Defines score interpretation and SQL thresholding, but stops short of explaining the sentiment model’s internal operation.
- Examples
- Uses bakery delivery reviews and example phrases about excellent service, stale bagels, and an ordinary bakery visit.
- What you can do after
- Apply it via Snowflake Cortex and SQL to score reviews, assign classes, filter feedback, and track sentiment over time.
Read how Snowflake Data Engineering explains sentiment analysis →
Investing for Programmers
Sentiment analysis is a computational technique for determining attitudes, opinions, or emotional polarity expressed in text.
- How it is used
- Converts qualitative language into signals for comparing companies, monitoring reputation, and assessing nonfinancial risk.
- How deep it goes
- Connects sentiment analysis to transcripts, keywords, and risk judgments but stops short of explaining model mechanics or implementation.
- Examples
- Uses earnings-call transcripts, social-media posts, political speeches, legal-case coverage, and news sentiment.
- What you can do after
- Apply it to compare textual sentiment and investigate possible reputation or political risks while treating conclusions cautiously.
Read how Investing for Programmers explains sentiment analysis →
Software Testing with Generative AI
Sentiment analysis is the computational classification or scoring of text according to expressed sentiment or evaluative polarity.
- How it is used
- Evaluates whether fine-tuned model outputs align with expected outputs during inference testing.
- How deep it goes
- Explains automated comparison and score interpretation, then stops short of replacing manual human validation.
- Examples
- Uses held-out or new instruction-output data and compares each returned response with its expected response.
- What you can do after
- Diagnose tuning quality via sentiment scores, then select future fine-tuning or human-review steps.
Read how Software Testing with Generative AI explains sentiment analysis →
Spring AI in Action
Sentiment analysis is the computational classification of text according to expressed opinions or emotional polarity.
- How it is used
- Uses user sentiment to guide responses and smooth negative experiences in online and natural-language applications.
- How deep it goes
- Explains prompt design, data types, controller wiring, and endpoint results, but defers sentiment-model internals.
- Examples
- Evaluates increasingly contextual phrases about the end of the world, plus a positive lyric-like passage.
- What you can do after
- Build a Spring endpoint that returns sentiment polarity and an explanation for submitted text.
Hugging Face in Action
Sentiment analysis is the computational classification of text according to the expressed attitude or polarity, such as positive or negative sentiment.
- How it is used
- Applies pretrained text-classification models to detect positive or negative opinions in reviews and other paragraphs.
- How deep it goes
- Explains pipeline use, then descends through tokenization, attention masks, logits, class selection, and fine-tuning datasets.
- Examples
- Uses IMDb movie reviews, a positive movie sentence, and labeled Yelp restaurant and service reviews.
- What you can do after
- Builds a classifier with a pipeline, or operates a tokenizer and model directly to produce sentiment predictions.
Read how Hugging Face in Action explains sentiment analysis →
Related concepts
- classifierkind of · A classifier is a predictive computational model that assigns inputs to one or more discrete classes or categories.
- language modeluses · A language model is a computational model that represents patterns in linguistic tokens to estimate sequence probabilities and predict or generate language from context.
- fine-tuningused by · Fine-tuning is adapting a pretrained machine-learning model by training it further on targeted data or objectives for a task, domain, or behavior.
- ChatGPTuses · ChatGPT is OpenAI’s conversational AI service, which lets users interact with generative AI models and related tools through natural-language prompts.
- natural language processingkind of · Natural language processing is the field of computing concerned with computationally analyzing, interpreting, transforming, or generating human language.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.