book

Hands-On APIs for AI and Data Science

by Ryan Day

March 2025

Intermediate to advanced

352 pages

8h 51m

English

O'Reilly Media, Inc.

Read now

Unlock full access

Includes

Includes Quizzes

Has Sandbox

Why Should You Read This Book?Who This Book Is ForData ScientistsAPI Developers and DesignersJob Seekers and Role ChangersCreating Portfolio ProjectsUsing This BookWhat This Book Is NotWhy Fantasy Football?Get More Tips on APIs, AI, and Data ScienceConventions Used in This BookUsing Code ExamplesO’Reilly Online LearningHow to Contact UsAcknowledgments
How Do Data Scientists Use APIs?What Tools Do Data Scientists Use?Designing APIs for Data ScientistsIntroducing Your Part I Portfolio ProjectEvery API Has a StoryMeeting Your Company: SportsWorldCentralSWC Needs an APISelecting the First API ProductsIdentifying Potential UsersCreating User StoriesAdditional ResourcesSummary
API Architectural StylesRepresentational State Transfer (REST)Graph Query Language (GraphQL)gRPCYour Choice: RESTTechnology ArchitectureSoftware Used in This ChapterPythonGitHubGetting Started with Your GitHub CodespaceCreating Your GitHub AccountCloning the Part I RepositoryLaunching Your GitHub CodespaceTouring Your New CodespaceMaking Your First CommitAdditional ResourcesSummary
Components of Your APISoftware Used in This ChapterSQLiteSQLAlchemypytestCreating Your SQLite DatabaseCreating Database TablesUnderstanding Table StructureLoading Your DataAccessing Your Data Using PythonInstalling SQLAlchemy in Your EnvironmentCreating Python Files for Database AccessCreating the Database Configuration FileCreating SQLAlchemy Helper FunctionsInstalling pytest in Your EnvironmentTesting Your SQLAchemy CodeAdditional ResourcesSummary
Continuing Your Portfolio ProjectSoftware Used in This ChapterFastAPIHTTPXPydanticUvicornCopying Files from Chapter 3Installing the New Libraries in Your CodespaceCreating Python Files for Your APICreating Pydantic SchemasCreating Your FastAPI ControllerTesting Your APILaunching Your APIAdditional ResourcesSummary
Sending a Signal of TrustMaking Great API DocsCore FeaturesExtra FeaturesReviewing Examples of API DocumentationSleeper AppMyFantasyLeagueYahoo! Fantasy FootballViewing Your API’s Built-in DocumentationCopying Files from Chapter 4Documentation Option 1: Swagger UIDocumentation Option 2: RedocWorking with Your OpenAPI Specification FileContinuing Your Portfolio ProjectAdding Details to the OAS info ObjectAdding Tags to Categorize Your PathsAdding More Details to Individual EndpointsAdding Parameter DescriptionsViewing the Changes in Swagger UIRegression-Testing Your APIUpdating Your README.mdAdditional ResourcesSummary
Benefits and Responsibilities of Cloud DeploymentBenefitsResponsibilitiesChoosing a Cloud Host for Your ProjectSetting Up Your Project DirectoryUsing GitHub Codespaces as a Cloud HostDeploying to RenderSigning Up for RenderCreating a New Web ServiceAuto-Deploying a Change to Your APIShipping Your Application in a Docker ContainerVerifying Docker InstallationCreating a DockerfileCreating a .dockerignore FileBuilding a Container ImageRunning Your Container Image LocallyDeploying to AWSCreating a Lightsail Container ServiceInstalling the AWS CLIInstalling the Amazon Lightsail Container Services Plug-inConfiguring Your Login CredentialsPushing Your Container Image to LightsailCreating a Lightsail DeploymentUpdating Your API DocumentationAdditional ResourcesSummary
SDKs Bridge the GapPicking a Language for Your SDKStarting with a Minimum Viable SDKExpert Tip: Making Your SDK Easy to InstallExpert Tip: Making the SDK Consistent and IdiomaticBuilding a Feature-Rich SDKExpert Tip: Using Sane DefaultsExpert Tip: Providing Rich FunctionalityExpert Tip: Performing LoggingExpert Tip: Hiding Your API’s Complicated DetailsExpert Tip: Supporting Bulk DownloadsExpert Tip: Documenting Your SDKTesting Your SDKExpert Tip: Supporting Every Task the API SupportsCompleting Your Part I Portfolio ProjectAdditional ResourcesSummary

Using a Variety of API StylesHTTP BasicsHow to Consume APIs ResponsiblySeparation of Concerns: Using SDKs or Creating API ClientsHow to Build APIsHow to Test APIsAPI Deployment and ContainerizationUsing Version ControlIntroducing Your Part II Portfolio ProjectGetting Started with Your GitHub CodespaceCloning the Part II RepositoryLaunching Your GitHub CodespaceRunning the SportsWorldCentral (SWC) API LocallyAdditional ResourcesSummary
Custom Metrics for Sports AnalyticsUsing APIs as Data Sources for Fantasy Custom MetricsCreating a Custom Metric: The Shark League ScoreSoftware Used in This ChapterhttpxJupyter NotebookspandasInstalling the New Libraries in Your CodespaceLaunching Your API in CodespacesCreating an API Client FileCreating Your Jupyter NotebookAdding General Configuration to Your NotebookWorking with Your API DataCalculating the League Balance ScoreCalculating the League Juice ScoreCreating the Shark League ScoreAdditional ResourcesSummary
Types of Data Sources for Data PipelinesPlanning Your Data PipelineOrchestrating the Data Pipeline with Apache AirflowInstalling Apache Airflow in GitHub CodespacesCreating Your Local Analytics DatabaseLaunching Your API in CodespacesConfiguring Airflow ConnectionsCreating Your First DAGCoding a Shared FunctionRunning Your DAGSummary
Engaging Users with Interactive VisualizationsSoftware Used in This Chapternfl_data_pyStreamlitInstalling Streamlit and nfl_data_pyLaunching Your API in CodespacesReusing the Chapter 9 API Client FileCreating Your Streamlit AppUpdating the Entrypoint FileRunning Your Streamlit AppCreating the Team Rosters PageCreating the Team Stats PageDeploying Your Streamlit AppCompleting Your Part II Portfolio ProjectAdditional ResourcesSummary
The Overlap of AI and APIsDesigning APIs to Use with Generative AI and LLMsDefining Artificial IntelligenceGenerative AI and Large Language Models (LLMs)Creating Agentic AI ApplicationsIntroducing Your Part III Portfolio ProjectGetting Started with Your GitHub CodespaceCloning the Part III RepositoryLaunching Your GitHub CodespaceAdditional ResourcesSummary
Training Machine Learning ModelsNew Software Used in This ChapterONNX Runtimescikit-learnsklearn-onnxInstalling the New Libraries in Your CodespaceUsing the CRISP-DM ProcessBusiness UnderstandingData UnderstandingData PreparationModelingEvaluationDeploymentAdditional ResourcesSummary
Calling AI Using APIs (via LangChain)Creating a LangGraph AgentSigning Up for AnthropicLaunching Your GitHub CodespaceInstalling the New Libraries in Your CodespaceCreating Your Jupyter NotebookChatting with the LangGraph AgentRunning the SportsWorldCentral (SWC) API LocallyInstalling the swcpy Software Development Kit (SDK)Creating a LangChain ToolkitCalling APIs Using AI (with LangGraph)Chatting with Your Agent (with Tools)Additional ResourcesSummary
Architecture of Your ApplicationGetting Started with ChatGPTCreating a Custom GPTLaunching Your GitHub CodespaceRunning the SportsWorldCentral (SWC) API in GitHub CodespacesAdding the Servers Section to Your OAS FileCreating a GPT ActionTesting the APIs in Your GPTChatting with Your Custom GPTCompleting Your Part III Portfolio ProjectSummary

Overview

Are you ready to grow your skills in AI and data science? A great place to start is learning to build and use APIs in real-world data and AI projects. API skills have become essential for AI and data science success, because they are used in a variety of ways in these fields. With this practical book, data scientists and software developers will gain hands-on experience developing and using APIs with the Python programming language and popular frameworks like FastAPI and StreamLit.

As you complete the chapters in the book, you'll be creating portfolio projects that teach you how to:

Design APIs that data scientists and AIs love
Develop APIs using Python and FastAPI
Deploy APIs using multiple cloud providers
Create data science projects such as visualizations and models using APIs as a data source
Access APIs using generative AI and LLMs