MISSION BRIEF
From public data to an interactive analytics experience.
This project turns North Carolina industry employment and wage data into three interactive visualizations, supported by transparent methodology and an editorial research narrative. The case study documents the application architecture, analytical decisions, AI-assisted development, and production validation.
Tools: Next.js · React · TypeScript · Apache ECharts · Sanity CMS · ChatGPT · GitHub · Vercel
THE DATA PRODUCT
Explore the published interactive dashboard.
Open the companion inquiry to explore the visualizations and analytical findings. The technical implementation is documented below.
Open the interactive dashboard →TECHNICAL FIELD LOG
Architecture, methodology, and implementation.
Project Overview
Project: North Carolina Industry Growth Dashboard
Period analyzed: 2019–2025
Geographic scope: North Carolina
Primary data sources: U.S. Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW) and Consumer Price Index for All Urban Consumers (CPI-U)
Technology stack: Next.js, React, TypeScript, Apache ECharts, Papa Parse, Sanity CMS, GitHub, and Vercel
Business and Analytical Challenge
Employment growth is frequently used as an indicator of economic strength. However, an industry adding jobs does not necessarily deliver meaningful improvements in workers' purchasing power.
This project explores two related questions:
Which North Carolina industries experienced the strongest employment growth between 2019 and 2025?
Did average weekly wage growth keep pace with inflation during the same period?
Answering these questions requires more than comparing employment totals. It requires consistent industry classifications, comparable reporting periods, reliable growth calculations, and a distinction between nominal wages and inflation-adjusted purchasing power.
Project Objective
I developed an interactive analytics application that allows readers to examine employment and wage trends across North Carolina industries.
The solution combines three complementary visualizations:
- Employment Growth: Compares percentage employment growth and absolute jobs added.
- Wage Growth: Compares nominal and inflation-adjusted average weekly wage growth.
- Employment vs. Wage Growth: Examines how workforce expansion relates to wage performance across industries.
These visualizations are supported by an editorial research narrative, three analytical finding cards, and documentation of the underlying methodology.
Intended Audience and Decision Support
The dashboard is designed for readers interested in North Carolina's economic development, labor market conditions, and industry performance.
It supports exploratory questions such as:
- Which industries are expanding fastest relative to their initial size?
- Which industries are contributing the largest absolute employment gains?
- Where are average weekly wages increasing faster than inflation?
- Which industries are adding jobs while experiencing declining inflation-adjusted wages?
The application supports comparative analysis and research rather than forecasting or making prescriptive employment or investment recommendations.
Solution Delivery
The dashboard was developed using Next.js, React, TypeScript, and Apache ECharts. Sanity CMS manages the accompanying editorial content, while GitHub and Vercel support source control and production deployment.
ChatGPT assisted with implementation, troubleshooting, and documentation. I retained responsibility for the research questions, analytical definitions, interpretation, review, and publication decisions.
The finished solution was published as an interactive inquiry on Field Notes & Data.
Technical Architecture
The North Carolina Industry Growth Dashboard was developed as an interactive analytics application using Next.js, React, TypeScript, and Apache ECharts.
The architecture separates analytical data, data transformation, visualization configuration, chart rendering, and editorial content. This approach supports reusable components and allows the published research narrative to be managed independently of the application code.
Application Framework
Next.js provides the application framework, while React manages interactive components and application state. TypeScript provides static typing for data structures, chart configurations, and component properties.
The application uses client-side React components for interactive visualizations, allowing users to change metrics, sort results, and explore industry comparisons without navigating away from the page.
Data Retrieval and Transformation
The dashboard retrieves a cleaned CSV dataset derived from the U.S. Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW).
The employment growth component uses the browser Fetch API to retrieve the dataset and Papa Parse to convert CSV records into structured objects.
Additional client-side processing applies analytical rules:
- Restricts records to the selected ownership category and NAICS industry-sector level.
- Selects observations from 2019 and 2025.
- Excludes suppressed records and invalid employment values.
- Matches industry observations across years using industry codes.
- Excludes industries without valid observations in both comparison years.
- Calculates absolute employment changes and percentage employment growth.
This separation between the prepared dataset and client-side transformations enables interactive analytical comparisons.
Visualization Architecture
The dashboard uses three specialized React visualization components:
- Employment Growth: Compares percentage employment growth and absolute jobs added, with dynamic ranking and sortable tabular results.
- Wage Growth: Compares nominal and inflation-adjusted wage growth.
- Employment vs. Wage Growth: Presents the relationship between employment and wage changes through a scatter plot.
Each visualization uses a shared EChartsChart component responsible for rendering Apache ECharts configurations.
The shared component uses Next.js dynamic imports with server-side rendering disabled, accommodating the charting library's browser-dependent functionality.
TypeScript's EChartsOption type defines chart configurations, while configurable component properties support consistent rendering across the application.
React's useMemo is used in the employment growth component to derive ranked and sorted datasets when relevant state changes.
Content Management
Sanity CMS manages the editorial content accompanying the interactive dashboard, including the research narrative, analytical findings, methodology, data sources, and documentation of AI-assisted development.
Separating editorial content from the visualization code allows the research narrative to be updated without modifying the chart components.
Deployment and Validation
GitHub provides source control, and Vercel hosts the production Next.js application.
The release workflow includes local development, production build verification, TypeScript validation, Git commits, Vercel deployment, and testing of the published interactive visualizations.
This process combines application development with analytical review and production verification.
Data Preparation and Analytical Methodology
Data Sources
The analysis uses publicly available economic data from the U.S. Bureau of Labor Statistics (BLS):
- Quarterly Census of Employment and Wages (QCEW): Industry-level employment and average weekly wage data for North Carolina.
- Consumer Price Index for All Urban Consumers (CPI-U): U.S. city average, all-items annual index values used to adjust wages for inflation.
The analysis compares 2019 and 2025 to evaluate changes in industry employment and average weekly wages over the six-year period.
Data Preparation and Quality Controls
The dashboard uses a cleaned CSV dataset derived from BLS QCEW data. Each visualization retrieves the dataset and applies additional filtering and analytical transformations in the browser.
The application applies the following data-quality rules:
- Restricts observations to ownership code 5 and the NAICS sector classification level.
- Includes records from 2019 and 2025.
- Excludes records marked as suppressed.
- Excludes missing, nonnumeric, zero, or negative values for the measures being analyzed.
- Matches observations across years using industry codes.
- Excludes industries without valid observations for both comparison years.
The employment growth visualization requires valid employment observations in both years. The wage growth visualization requires valid average weekly wage observations in both years.
The employment-versus-wage-growth scatterplot requires both measures to be valid in both years. Consequently, the set of industries included in the scatterplot may differ from those in the individual charts.
These validation rules support consistent year-over-year comparisons and prevent invalid or incomplete records from entering the displayed calculations.
Analytical Measures
The dashboard calculates three primary growth measures.
1. Employment Growth
Employment growth measures the percentage change in annual average employment between 2019 and 2025.
Employment Growth (%) = ((2025 Employment / 2019 Employment) − 1) × 100
The dashboard also calculates absolute employment change:
Jobs Added = 2025 Employment − 2019 Employment
These measures distinguish industries experiencing rapid percentage growth from those contributing the largest number of additional jobs.
2. Nominal Wage Growth
Nominal wage growth measures the percentage change in average weekly wages without adjusting for inflation.
Nominal Wage Growth (%) = ((2025 Average Weekly Wage / 2019 Average Weekly Wage) − 1) × 100
3. Inflation-Adjusted Wage Growth
To account for changes in purchasing power, 2019 average weekly wages are converted into equivalent 2025 dollars using CPI-U.
2019 Wage in 2025 Dollars = 2019 Wage × (2025 CPI-U / 2019 CPI-U)
Real Wage Growth (%) = ((2025 Wage / 2019 Wage in 2025 Dollars) − 1) × 100
The dashboard implementation uses annual CPI-U values of 255.657 for 2019 and 321.943 for 2025, producing an inflation adjustment factor of approximately 1.2593.
This adjustment allows nominal wage changes to be distinguished from changes in inflation-adjusted purchasing power.
Analytical Interpretation
The distinction between nominal and real wage growth is essential. An industry can experience rising average weekly wages in nominal dollars while delivering much smaller gains—or declines—in purchasing power after accounting for inflation.
The employment-versus-wage-growth scatterplot combines two dimensions of industry performance. Employment growth appears on the horizontal axis, while the vertical axis can display either nominal or real wage growth.
This enables comparisons of industries experiencing workforce expansion, contraction, wage gains, or inflation-adjusted wage declines.
The employment chart additionally allows comparisons using either percentage employment growth or absolute jobs added, providing different perspectives on industry expansion.
Limitations
The analysis describes aggregate industry-level changes rather than the experiences of individual workers. Changes in workforce composition, industry classifications, hours worked, and employment patterns may influence observed average weekly wages.
Average weekly wage growth should not be interpreted as the salary increase received by a typical individual employee.
The inflation adjustment uses national CPI-U values rather than a North Carolina-specific measure of living costs.
The comparisons are descriptive and do not establish that employment growth caused wage growth or vice versa.
Results should be interpreted within the geographic scope, industry classifications, reporting periods, and coverage of the underlying BLS datasets. Differences in data availability and validation requirements may also affect which industries appear in each visualization.
Interactive Visualization Design
Design Objectives
The dashboard was designed to turn industry-level economic statistics into an accessible analytical experience without sacrificing the detail required for meaningful comparisons.
Three design principles guided the visualizations:
- Comparative clarity: Make differences between industries easy to identify.
- Analytical flexibility: Allow readers to switch measures and explore alternative rankings.
- Transparent interpretation: Connect the visualizations to clearly defined metrics, source data, and documented limitations.
Rather than relying on a single chart, the dashboard presents three complementary views that address different analytical questions.
Employment Growth — Relative vs. Absolute Expansion
The employment growth visualization compares North Carolina industries using two measures: percentage employment growth and absolute jobs added.
Percentage growth normalizes changes relative to an industry's initial employment level. This helps readers identify sectors expanding rapidly even when they began with relatively small workforces.
Absolute jobs added provides a different perspective by showing the number of additional positions associated with each industry.
The visualization supports:
- Switching between percentage growth and jobs added.
- Dynamically ranking industries according to the selected measure.
- Reviewing underlying employment values in an interactive table.
- Sorting the table by industry name, employment totals, jobs added, or percentage growth.
The chart also accommodates horizontal scrolling on smaller screens.
Design rationale: A percentage-only ranking can overemphasize smaller industries, while an absolute-growth ranking can favor larger industries. Providing both perspectives helps readers evaluate relative performance and the scale of employment change.
Wage Growth — Nominal vs. Inflation-Adjusted Performance
The wage growth visualization examines changes in average weekly wages between 2019 and 2025.
Readers can switch between nominal and inflation-adjusted wage growth, with the industry ranking updating to reflect the selected measure.
This interaction addresses a common analytical problem: increases in wages measured in current dollars do not necessarily indicate improvements in purchasing power.
The supporting data table provides additional context through wage values and calculated growth measures.
Design rationale: Keeping nominal and real wage growth within the same visualization makes the distinction between the two measures central to the analytical experience, rather than treating inflation as a secondary footnote.
Employment vs. Wage Growth — Multivariable Analysis
The scatterplot combines employment growth and wage growth in a single analytical view.
Each industry is represented using:
- Horizontal axis: Percentage employment growth.
- Vertical axis: Percentage growth in average weekly wages, selectable between nominal and inflation-adjusted measures.
This design allows readers to examine different combinations of employment and wage performance.
For example, an industry may experience substantial employment growth while average weekly wages decline after inflation. Another industry may experience relatively limited employment growth alongside stronger wage gains.
Interactive tooltips provide additional industry-level information, including employment growth, nominal wage growth, real wage growth, and jobs added.
Design rationale: The scatterplot makes it possible to explore relationships, contrasts, and potential outliers that are less apparent when employment and wage performance are viewed separately.
The visualization is exploratory. Its patterns should not be interpreted as evidence of causation.
Guided Storytelling and Independent Exploration
The dashboard combines three editorial finding cards with interactive visualizations.
The finding cards communicate important observations and provide readers with a starting point for understanding the results.
The interactive charts then allow readers to investigate industries and comparisons beyond those emphasized in the narrative.
This creates two complementary paths through the analysis:
- Guided analysis: Readers can understand the main findings without manipulating the dashboard.
- Independent exploration: Readers can change measures, review industry-level values, and examine alternative comparisons.
This approach combines the accessibility of editorial storytelling with the flexibility of a self-service analytics experience.
Visualization Implementation
Apache ECharts provides the charting capabilities within the Next.js and React application.
Each visualization component manages its own analytical data and chart configuration, while a shared rendering component provides consistent integration with ECharts.
The employment and wage charts use React state to manage metric selections and sorting behavior. Memoized calculations support the derivation of chart rankings and table views from the loaded dataset.
The visualizations are embedded within the published research narrative, connecting the analytical results to the methodology and interpretation.
The overall design emphasizes meaningful comparisons and user-directed exploration rather than visual complexity for its own sake.
AI-Assisted Development and Human Validation
Development Approach
AI-assisted development was incorporated into the project as a practical method for accelerating implementation, troubleshooting, and documentation.
ChatGPT supported development activities involving Next.js, React, TypeScript, Apache ECharts, and Sanity CMS.
The workflow was iterative: implementation guidance was generated, applied to the local project, reviewed, tested, and refined as issues or new requirements emerged.
AI assistance supported the development process but did not replace the need to understand the analytical requirements or validate the resulting application.
Division of Responsibilities
The project combined AI-assisted implementation with human-directed analytical and design decisions.
AI-assisted activities included:
- Developing and refining application code.
- Supporting React and TypeScript implementation.
- Assisting with visualization configuration and interactive features.
- Troubleshooting development and deployment issues.
- Supporting technical documentation and editorial revisions.
Human-directed responsibilities included:
- Defining the research questions and analytical objectives.
- Selecting the employment and wage measures.
- Determining how inflation should be incorporated into the comparisons.
- Reviewing the analytical calculations and data presentation.
- Evaluating visualization usability and interpretability.
- Testing the application and deciding when it was ready for publication.
This division of responsibilities is important because generated code can be syntactically plausible while still containing incorrect assumptions, unsuitable analytical choices, or implementation errors.
Analytical Validation
Validation focused on whether the dashboard implemented the intended analytical definitions and produced interpretable results.
The review process included examining employment and wage growth calculations, distinguishing nominal from inflation-adjusted wages, and checking the treatment of incomplete or invalid records.
The application's data-processing logic applies explicit validation rules, including checks for suppressed observations, valid numeric values, consistent industry classifications, and comparable reporting years.
These rules are important because errors in source-data selection or metric definitions can undermine an otherwise polished visualization.
The validation described here represents implementation review and functional testing; it should not be interpreted as an independently audited statistical analysis.
Application Testing and Deployment
The application underwent iterative testing during development and before publication.
The release workflow included:
- Testing the three interactive visualizations in the local development environment.
- Reviewing metric-switching, chart behavior, and analytical presentation.
- Running the Next.js production build and TypeScript checks.
- Committing application changes to GitHub.
- Deploying the application through Vercel.
- Verifying that unpublished editorial content was not publicly displayed.
- Publishing the research content through Sanity CMS.
- Testing the live dashboard and its accompanying research narrative.
Testing the production experience was particularly important because successful local development does not guarantee that chart rendering, data retrieval, or content publishing will behave identically after deployment.
Lessons from AI-Assisted Development
The project reinforced several principles for using generative AI in analytics development.
1. Define analytical requirements before implementation.
Clear definitions of employment growth, wage growth, and inflation adjustment provide a basis for evaluating generated implementation suggestions.
2. Treat generated code as a proposal, not proof of correctness.
AI-generated solutions require inspection, testing, and refinement before they can be trusted in a published analytics application.
3. Preserve human accountability for interpretation.
AI can support the development of analytical tools, but the developer remains responsible for selecting measures, understanding limitations, and communicating results accurately.
4. Validate the complete application.
Technical quality includes more than compiling code. Data retrieval, calculations, interactive behavior, editorial content, and production deployment must work together.
The resulting workflow demonstrates how AI-assisted software development can complement established business intelligence and analytical practices.
Project Outcomes and Lessons Learned
Delivered Solution
The project resulted in a publicly accessible, interactive analytics experience examining North Carolina industry employment and average weekly wage growth between 2019 and 2025.
The delivered solution includes:
- Three interactive visualizations covering employment growth, wage growth, and the relationship between the two measures.
- Interactive controls for comparing percentage growth, absolute jobs added, and nominal versus inflation-adjusted wages.
- Industry-level tabular information supporting additional exploration.
- Three editorial finding cards communicating key observations.
- A prepared analytical dataset derived from publicly available BLS statistics.
- Documentation of the data sources, methodology, limitations, AI assistance, and validation process.
- A production Next.js application deployed through Vercel, with source control in GitHub and editorial content managed through Sanity CMS.
The dashboard was published and tested as part of the Field Notes & Data website.
Analytical Value
The project demonstrates how multiple analytical perspectives can provide a more informative view of industry performance than employment growth alone.
Percentage employment growth identifies industries expanding rapidly relative to their initial size, while absolute jobs added reveals the scale of employment changes.
Nominal and inflation-adjusted wage measures distinguish increases in reported wages from changes in purchasing power.
The scatterplot combines employment and wage performance, enabling readers to identify industries where workforce expansion and wage growth move in different directions.
Together, these views support more nuanced exploration of North Carolina's industry-level economic changes.
Technical Outcomes
The project demonstrates the implementation of a complete interactive analytics workflow:
- Retrieving and parsing structured data in a browser-based application.
- Applying validation rules and calculating analytical measures.
- Developing typed React visualization components.
- Separating chart-specific logic from reusable rendering functionality.
- Supporting interactive metric selection, ranking, and tabular exploration.
- Integrating analytics components with CMS-managed editorial content.
- Deploying and validating a production web application.
The result is a working portfolio example connecting analytical methodology, interactive application development, and public-facing communication.
Technical Lessons Learned
1. Reusable components improve maintainability.
Separating specialized analytical components from shared chart-rendering functionality reduces duplication and establishes a clearer application structure.
2. Data quality is part of visualization engineering.
Validating source records, matching comparable observations, and excluding incomplete data are necessary steps before presenting analytical results.
3. Metric selection changes the interpretation.
Percentage growth and absolute change answer different questions. Nominal and real wage growth also provide different perspectives on economic performance.
Providing multiple measures helps users understand these distinctions rather than relying on a single ranking.
4. Analytical storytelling and self-service exploration complement each other.
Editorial findings help readers understand the primary observations, while interactive charts allow them to investigate additional questions independently.
5. AI assistance requires structured validation.
AI-supported development can accelerate implementation, but analytical accuracy, usability, and production reliability remain the responsibility of the developer.
6. Production verification is part of delivery.
A completed analytics application must function in its deployed environment, including data retrieval, interactive charts, and published editorial content.
Professional Application
This project demonstrates an end-to-end approach to delivering an interactive analytics solution, from framing the research questions and defining metrics to implementing visualizations, documenting methodology, and deploying a production application.
It also demonstrates how established business intelligence and data visualization practices can be extended through modern web technologies and AI-assisted development.
The most relevant transferable capabilities include analytical requirements definition, data validation, visualization design, interactive application development, technical documentation, and iterative delivery.
While the project is an independently developed portfolio application rather than a commissioned enterprise implementation, its development process reflects many of the same considerations involved in delivering professional analytics solutions: clear requirements, defensible calculations, maintainable components, meaningful user interactions, and production validation.
Explore the Published Inquiry
The companion inquiry presents the research findings and interactive dashboard.
Which North Carolina Industries Are Growing the Fastest—and Are Their Wages Keeping Up?
https://fieldnotesanddata.com/inquiries/north-carolina-industry-growth
AI WORKFLOW
How AI supported development.
ChatGPT served as an AI-assisted development partner throughout the project, providing implementation guidance, code suggestions, troubleshooting support, and technical documentation. AI assistance supported the development of the Next.js and React application, TypeScript components, Apache ECharts visualizations, and Sanity CMS integration. It also helped structure the development workflow into manageable steps and troubleshoot issues encountered during local testing and deployment. The project followed an iterative process in which AI-generated suggestions were reviewed, implemented, tested, and refined. ChatGPT accelerated development and supported work with unfamiliar technologies, but its outputs were not treated as independently validated. Research direction, analytical interpretation, visualization decisions, and final publication approval remained under human control.
HUMAN REVIEW
Analytical judgment and validation.
I retained responsibility for defining the research objectives, selecting the analytical comparisons, evaluating the visualizations, and determining how the findings should be communicated. My analytical decisions included comparing North Carolina industry employment and wage growth between 2019 and 2025, distinguishing nominal wage growth from inflation-adjusted wage growth, and using complementary visualizations to communicate different dimensions of industry performance. Throughout development, I reviewed AI-assisted implementation suggestions and tested the application's functionality rather than assuming generated code was correct. Validation activities included: - Reviewing employment growth and inflation-adjusted wage calculations. - Testing the employment growth, wage growth, and scatter plot visualizations. - Verifying that interactive chart controls behaved as intended. - Reviewing the editorial findings, methodology, and data source documentation. - Confirming that the Next.js production build and TypeScript checks passed. - Verifying successful deployment through GitHub and Vercel. - Confirming that unpublished Sanity content remained hidden from the public website. - Testing all three visualizations and the accompanying content after publication. I made the final decisions regarding analytical presentation, application readiness, and publication. AI assistance supported implementation, but responsibility for reviewing and approving the finished work remained mine.
REPRODUCIBILITY
Data sources and limitations.
BLS Quarterly Census of Employment and Wages (QCEW) — Primary source of North Carolina industry-level employment and wage data for the 2019–2025 comparison. Used to calculate employment growth, nominal wage growth, and industry-level comparisons across the dashboard.
BLS Consumer Price Index for All Urban Consumers (CPI-U) — Annual CPI-U values used to adjust historical wages for inflation. The analysis uses CPI-U values of 255.657 for 2019 and 321.943 for 2025 to express 2019 wages in 2025 dollars and calculate inflation-adjusted wage growth.
**Data Coverage and Comparability** The analysis uses publicly available U.S. Bureau of Labor Statistics data to compare North Carolina industry employment and wages between 2019 and 2025. Findings are limited to the geographic coverage, industry classifications, reporting periods, and employment coverage of the underlying QCEW dataset. **Industry-Level Aggregation** The analysis uses industry-level statistics rather than individual employee records. Changes in average wages may reflect changes in workforce composition, occupational mix, hours worked, or other factors rather than wage increases experienced by individual workers. **Inflation Adjustment** Real wage growth is calculated using national CPI-U values for 2019 and 2025. This provides a consistent inflation adjustment but does not account for differences in local living costs across North Carolina or individual household spending patterns. **Historical Comparison** The analysis compares two reference years rather than modeling every intervening period. As a result, the visualizations describe net changes between 2019 and 2025 but do not fully capture fluctuations during those years. **Correlation and Causation** The employment-versus-wage-growth scatter plot supports exploratory comparisons but does not establish a causal relationship between industry employment growth and wage growth. **Technical Scope** The dashboard uses a prepared CSV dataset rather than a live connection to BLS data services. Consequently, the visualizations do not automatically update when new economic data becomes available. **AI-Assisted Development** AI-generated implementation suggestions required human review and testing. Successful application deployment and functional testing do not independently guarantee the accuracy of every analytical assumption or interpretation. **Future Improvements** Potential enhancements include automated data refreshes, additional historical periods, more detailed industry breakdowns, and expanded validation of economic measures.
