Exercise

✅ Analyze-text-with-Azure-AI-Language

🧩 Problem You need to analyze text documents (e.g., reviews) to automatically extract insights such as the language, sentiment, key topics, entities, and linked entities to improve decision-making or content understanding.

💡 Solution with Azure Use the Text Analytics capabilities in Azure AI Language to analyze documents. It enables language detection, sentiment analysis, key phrase extraction, named entity recognition, and detection of linked entities via SDK or REST API.

🧱 Required Components Azure AI Language resource

Endpoint and subscription key (from Azure Portal)

Text Analytics SDK for C# or Python (azure-ai-textanalytics==5.3.0)

(Optional) .env or appsettings.json for secure key configuration

Cloud Shell + GitHub repo: https://github.com/microsoftlearning/mslearn-ai-language

🏗️ Architecture / Development 1. Provision Azure AI Language Resource In Azure Portal → Create resource → Search "Language Service"

Use Pricing tier F0 (free) if available

Once deployed, go to Keys and Endpoint to retrieve connection details

  1. Clone Code Repository in Azure Cloud Shell bash Copia Modifica rm -r mslearn-ai-language -f git clone https://github.com/microsoftlearning/mslearn-ai-language cd mslearn-ai-language/Labfiles/01-analyze-text
  2. Install SDK & Configure App Navigate to:

cd Python/text-analysis or cd C-Sharp/text-analysis

Python:

bash Copia Modifica python -m venv labenv ./labenv/bin/Activate.ps1 pip install -r requirements.txt azure-ai-textanalytics==5.3.0 C#:

bash Copia Modifica dotnet add package Azure.AI.TextAnalytics --version 5.3.0 4. Update Config File with Azure Credentials Python: edit .env

C#: edit appsettings.json

Add ai_key and ai_endpoint

💬 Text Analysis Functionalities a. Language Detection Detects the language of a text sample.

Python:

python Copia Modifica detectedLanguage = ai_client.detect_language(documents=[text])[0] print("Language:", detectedLanguage.primary_language.name) C#:

csharp Copia Modifica DetectedLanguage detectedLanguage = aiClient.DetectLanguage(text); Console.WriteLine($"Language: {detectedLanguage.Name}"); b. Sentiment Analysis Classifies the sentiment as positive, negative, neutral, or mixed.

Python:

python Copia Modifica sentimentAnalysis = ai_client.analyze_sentiment(documents=[text])[0] print("Sentiment:", sentimentAnalysis.sentiment) C#:

csharp Copia Modifica DocumentSentiment sentimentAnalysis = aiClient.AnalyzeSentiment(text); Console.WriteLine($"Sentiment: {sentimentAnalysis.Sentiment}"); c. Key Phrase Extraction Extracts main concepts or ideas from the text.

Python:

python Copia Modifica phrases = ai_client.extract_key_phrases(documents=[text])[0].key_phrases for phrase in phrases: print(phrase) C#:

csharp Copia Modifica KeyPhraseCollection phrases = aiClient.ExtractKeyPhrases(text); foreach (string phrase in phrases) Console.WriteLine(phrase); d. Named Entity Recognition (NER) Identifies named entities like locations, people, organizations.

Python:

python Copia Modifica entities = ai_client.recognize_entities(documents=[text])[0].entities for entity in entities: print(f"{entity.text} ({entity.category})") C#:

csharp Copia Modifica CategorizedEntityCollection entities = aiClient.RecognizeEntities(text); foreach (var entity in entities) Console.WriteLine($"{entity.Text} ({entity.Category})"); e. Linked Entity Recognition Finds entities with external links (e.g., Wikipedia).

Python:

python Copia Modifica entities = ai_client.recognize_linked_entities(documents=[text])[0].entities for entity in entities: print(f"{entity.name} ({entity.url})") C#:

csharp Copia Modifica LinkedEntityCollection linkedEntities = aiClient.RecognizeLinkedEntities(text); foreach (var linkedEntity in linkedEntities) Console.WriteLine($"{linkedEntity.Name} ({linkedEntity.Url})"); ✅ Best Practices / Considerations Use Cloud Shell for quick development/testing

Use .env/appsettings.json to secure sensitive info

Batch process documents to reduce API calls

Analyze sentiment trends over time in reviews or feedback

Detect named and linked entities for deeper contextual understanding

Consider rate limits and pricing tier when analyzing large text volumes