What you will learn in Data Storytelling with Altair and AI
Chapter by chapter
Concepts explained in this book
- ChatGPTChatGPT is a generative AI conversational system that produces and transforms text from user prompts.
- promptA prompt is an input text that provides instructions to a generative AI system about the content or operation to produce.
- data storytellingA data story is a narrative that communicates insights by integrating data, visual representations, and explanatory context.
- contextTextual context is written material that provides background or explanation around a data visualization.
- informationInformation is data presented in a form that conveys meaning but may still lack sufficient surrounding context.
- target audienceAn audience is the person or group intended to read, interpret, or receive a data story or presentation.
- mark_text()mark_text() is an Altair method that creates text marks for displaying textual content in a visualization.
- insightAn insight is a meaningful conclusion or understanding derived from interpreting data.
- AltairAltair is a declarative Python library for statistical visualization based on the Vega-Lite visualization grammar.
- DALL-EA generative AI system that creates and edits images from textual descriptions or image-editing instructions.
- dataData is a collection of recorded observations, measurements, or other representations of facts before interpretation.
- encoding channelAn encoding channel is a visual dimension that maps a data field to a graphical property such as position or color.
- retrieval-augmented generationRetrieval-augmented generation is a technique that supplies a language model with relevant information retrieved from an external knowledge source when generating output.
- wisdomWisdom is knowledge enriched with ethics that supports action and suggests subsequent steps.
- call-to-actionA call to action is an explicit invitation for an audience to engage in a specified behavior or event.
- DIKW hierarchyThe DIKW pyramid is a conceptual hierarchy that organizes data, information, knowledge, and wisdom as progressively interpreted forms of understanding.
- percentage changeA percentage increase is a measure of relative growth calculated by dividing the change in a quantity by its original value and expressing the result as a percentage.
- encodingAn encoding is a mapping that assigns data fields to visual properties of chart marks.
- hallucinationA hallucination is a generative AI output that convincingly resembles valid or real content despite lacking grounding in reality or the underlying source.
- mark_line()mark_line() is an Altair method that creates line marks for connecting data values in a visualization.
- user interactionChart interactivity is the ability of a visualization to respond to user input such as hovering, clicking, zooming, or panning.
- fine-tuningFine-tuning is the additional training of a pretrained machine-learning model on targeted data to adapt its parameters to a task or domain.
- base chartA base chart is a chart specification containing shared encodings and properties from which related charts are constructed.
- bar chartA bar chart is a chart that represents values with rectangular bars whose lengths or heights encode magnitude.
- mark_image()mark_image() is an Altair API method that specifies image marks for displaying image URLs or image resources in a chart.
- visual markA mark is a graphical primitive in a visualization grammar that represents data items through visual shapes such as points, lines, or bars.
- data frameA pandas DataFrame is a two-dimensional labeled tabular data structure with rows and columns.
- biasBias is a systematic distortion or inclination that influences decisions, perceptions, judgments, or representations in favor of or against a particular subject.
- VegaA Vega specification is a JSON document that defines a chart's data, visual appearance, and interactive behavior.
- tourist arrivalsTourist arrivals are recorded entries or visits by tourists to a destination or accommodation establishment during a specified period.
- input-output pairAn input-output pair is a training example consisting of an input instruction or prompt and the corresponding desired model response.
- JSONLJSONL is a line-oriented text format in which each line contains a separate JSON value, commonly an object.
- manual dataset buildingManual dataset building is the deliberate creation of training examples by writing each prompt and desired completion individually.
- source-based dataset buildingSource-based dataset building is the creation of training data by extracting examples from external materials such as websites, blogs, or scientific papers.
- language modelA large language model is a neural language model with many learned parameters that generates or transforms text based on input context.
- ethical frameworkAn ethical framework is a structured set of principles used to evaluate actions and guide decisions according to values about right and wrong.
- missing valueA missing value is an absent or unrecorded value for an observation and variable in a dataset.
- human emotionsAn emotion is a feeling or affective state represented or conveyed by an image or visual element.
- slope graphA slope graph, or slope chart, is a line-based chart that compares values at two endpoints by connecting corresponding points.
- moodMood is the prevailing emotional tone or affective valence conveyed by a communication or experienced by its audience.
- vector databaseA vector database is a database optimized for storing and retrieving numerical vector representations by similarity or other vector operations.
- transform_calculate()transform_calculate() is an Altair chart method that derives and adds fields using expression-based transformations.
- colorColor is a visual attribute describing the hues or tonal properties represented in an image or visualization.
- configure_axis()configure_axis() is an Altair chart method that sets default properties shared by chart axes.
- Vega-LiteVega-Lite is a high-level declarative language built on Vega for specifying common data visualizations with concise syntax.
- embeddingAn embedding is a numerical vector representation that captures semantic or structural properties of an object such as text, enabling comparison and computation in a continuous vector space.
- machine learningMachine learning is a subfield of artificial intelligence in which computers learn patterns from examples or data to improve performance and make predictions or decisions.
- Chroma DBChroma is an embedding database for storing and retrieving vector representations of data.
- LangChainLangChain is a software framework for building applications that connect language models with prompts, external data sources, and other components.
- templateA prompt template is predefined text containing instructions and variable placeholders for constructing prompts for language models.
- information retrievalInformation retrieval is the process of searching for and returning relevant information from a collection or external knowledge source.
- mark_rule()mark_rule() is an Altair API method that specifies rule marks, such as horizontal or vertical reference lines.
- compound chartA compound chart is a visualization composed of multiple charts combined into a coordinated graphical display.
- imperative visualizationImperative visualization is a programming approach in which code specifies the individual steps for constructing a visualization.
- axes sectionAn axes section is the part of a Vega specification that defines chart axes, including their scales, orientations, tick marks, and labels.
- data sectionA data section is the part of a Vega specification that defines data sources, formats, locations, loading, and transformations.
- marks sectionA marks section is the part of a Vega specification that defines graphical marks, their data source, encoding, and visual styling.
- scales sectionA scales section is the part of a Vega specification that defines mappings from data domains to visual ranges.
- signals sectionA signals section is the part of a Vega specification that declares interactive signals and the events that update them.
- baselineA baseline is a reference value or visual reference line used to compare observed values or improvements.
- stacked bar chartA stacked bar chart displays categorical quantities as segments within bars so component values form a total.
- digital imageAn image is a visual representation stored or rendered as graphical content.
- topicA topic is the subject of a chart that a prompt directs a generative AI system to address.
- good judgementGood judgment is the ability to evaluate a situation from multiple perspectives, weigh risks and benefits, and choose an informed course of action.
- RectangleA rectangle is a four-sided graphical shape used as a visual mark to cover a region of a chart.
- mark_rect()mark_rect() is an Altair API method that creates rectangular marks in a chart.
- data storytelling arcThe data storytelling arc is a narrative structure adapted for data stories that establishes background, develops audience interest, reveals a main point, and leads toward resolution or action.
- chart titleA chart title is textual content displayed with a chart to identify or describe its subject.
- data analysisThe process of examining, transforming, and interpreting data to produce findings or insights.
- conditionA condition in a visualization is a criterion-dependent rule that determines values or visual properties for different data cases.
- nominal dataNominal data contains categorical labels whose values have no intrinsic ordering.
- quantitative dataQuantitative data contains numerical values that represent amounts or measurements.
- configure_*()configure_view() is an Altair chart method that sets view-level configuration properties for single or compound charts.
- artificial intelligenceArtificial intelligence is the field of computing concerned with systems that perform tasks requiring capabilities associated with human intelligence.
- dropdownA dropdown menu is an interactive control that reveals a list of choices from which a user can select an option.
- configure_title()configure_title() is an Altair API method that sets global visual configuration for chart titles.
- href channelA URL channel maps data values to external resource locations associated with visual marks.
- PythonPython is a general-purpose, high-level programming language widely used for data analysis and scientific computing.
- taskA task is a prompt component that specifies the operation or result the generative AI system should perform or produce.
- mark functionmark_bar() is an Altair API method that specifies bar marks for a chart.
- subtitleA subtitle is a secondary heading that supplements or clarifies a title.
- TitleParams()TitleParams() is an Altair configuration object for specifying chart title text and related title properties.
- cooking chart familyA cooking chart family is a book-specific grouping of pie charts and donut charts for showing parts of a numerical whole.
- scaleA visualization scale maps data values to visual values such as positions, colors, or sizes.
- prompt structurePrompt structure is the organization of descriptive instructions that specify the intended content and visual characteristics of a generated image.
- CometComet is a machine-learning experiment management platform for tracking experiments, models, and evaluation results.
- emotional energyA high-energy emotion is an emotional state involving heightened physiological arousal, alertness, or activation.
- high-mood emotionA high-mood emotion is an emotional state associated with positive affective valence, happiness, or optimism.
- low-energy emotionA low-energy emotion is an emotional state involving reduced physiological arousal, calmness, or relaxation.
- low-mood emotionA low-mood emotion is an emotional state associated with negative affective valence, sadness, anger, or anxiety.
- layerA layer is a Vega-Lite composition that overlays multiple marks or charts within a single visualization.
- ratioA ratio is a quantitative comparison between two quantities obtained by division.
- independent scaleAn independent scale is a visualization scale resolved separately for one component of a composed chart rather than shared across components.
- lineA line is a graphical mark that connects plotted data positions to represent continuity or progression.
- resolve_scale()resolve_scale() is an Altair chart method that controls whether scales are shared or independent across composed charts.
- social media promotionSocial media promotion is the dissemination of information or marketing messages through social networking platforms.
- text channelA text channel maps data to textual content displayed as supplementary chart annotation.
- transform_filter()transform_filter() is an Altair API method that filters data rows within a chart specification.
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.