Projects
My work bridges art history, digital humanities, and computational analysis, using spatial mapping, data automation, and machine learning to uncover patterns in historical records. I specialize in digitizing archival sources, developing interactive visualizations, and analyzing the evolution of art markets and urban spaces. Through these projects, I aim to integrate technology with historical research, making complex data accessible and engaging.
Project name 01:
The Evolution Of The Art Market Reporting Through The Lens Of The New York Times, 1851-1900
Researcher
Number of The New York Times’ reports related to the art market over the years (1851-1900)
This thesis (View Here) examines the evolution of art market reporting in The New York Times from 1851 to 1900, tracing the increasing role of mainstream media in shaping public perceptions of the art market. While the dealer-critic system has been widely recognized in European contexts, the development of art market discourse in American newspapers remains underexplored. This study addresses this gap by analyzing 1,346 articles, identifying trends in how the press engaged with the art market during the second half of the nineteenth century.
Using keyword frequency analysis, this research tracks the growing presence of art market-related terminology over time, revealing a gradual increase in public interest and media attention. Additionally, the study examines four key art market events that influenced The New York Times’ coverage, demonstrating how shifts in reporting styles and editorial strategies reflected broader economic and cultural transformations.
By integrating historical analysis with computational text analysis, this thesis highlights how newspapers contributed to legitimizing the art market as a subject of mainstream discourse. It provides new insights into the intersections of journalism, economic history, and art criticism, emphasizing the press's role in shaping narratives about the commercial art world in nineteenth-century America.
Project name 02: Goupil & Co. in New York: Pioneer or Prudent Businessman?
Researcher
This study investigates Goupil & Cie's arrival in the New York art market in 1848, questioning whether the firm pioneered the city's art economy or responded to an already thriving market. While Goupil & Cie is often credited as a catalyst for New York's art scene, this research takes a closer look at what the market looked like before and after their arrival and whether their presence truly transformed it.
By exploring New York City Directories from 1843 to 1853, this study maps the key players in New York's art scene and examines how the market was growing.
Through GIS mapping and Kernel Density Estimation (KDE), the research highlights the growth and clustering of art-related businesses before and after Goupil’s arrival. The findings reveal a 43.65% increase in art-related occupations from 1842 to 1847, indicating a pre-existing and expanding market. Along with the quantitative analysis, the study considers how organizations like the American Art Union contributed to the market’s expansion, providing opportunities for artists and attracting new investors.
The research is brought to life through an interactive ArcGIS Story Map (View Here), offering a visual journey through the evolving art market of nineteenth-century New York. Rather than dramatically reshaping the scene, the findings suggest that Goupil & Cie leveraged an already growing art economy and adapted its business to fit this emerging landscape. This work provides a fresh perspective on how international art firms navigated the American market and what their presence meant for the future of art collecting and trade.
This project is currently featured in the 2025 CMAC Exhibition, Artificial Worlds: Creatures, Machines, and Perception, at Duke University. The exhibition runs from February 20 to March 5, 2025, at the Rubenstein Arts Center Gallery (Room 235, 2nd floor), with a closing reception on March 5 from 5:30 PM to 8:00 PM.
Project name 03: Virtual Black Charlotte
Graduate Research Assistant
The Virtual Black Charlotte project seeks to digitally reconstruct the historic Brooklyn neighborhood of Charlotte, NC, a once-thriving Black community (1951that was largely erased during urban renewal in the 1960s. Although much of its physical landscape has disappeared, Brooklyn remains an active site of memory for its former residents and their descendants. This project is a collaboration between Duke University, Johnson C. Smith University (JCSU), and the local Residents Council, with support from the National Archives. It aims to reclaim and preserve the history, culture and lived experiences of this lost neighborhood.
Rather than simply recreating buildings, Virtual Black Charlotte highlights the social and cultural fabric that made Brooklyn a vibrant hub of Black enterprise, faith, education, and community life. The project combines oral histories, archival records, historical maps, city directories, and architectural surveys to create a layered, immersive digital reconstruction. It also critically reflects on how digital methods and historical narratives shape our understanding of the past.
My Role in Virtual Black Charlotte
As part of Professor Victoria Szabo’s team, I contributed to three key areas of the project:
-
ArcGIS Mapping and Visualization:
-
Using ArcGIS, I helped visualize the spatial history of Brooklyn- tracing and Georeferencing streets and key locations. These interactive maps (View Here) allow audiences to explore Brooklyn’s past dynamically, providing insights into its urban structure, social spaces, and cultural landmarks before urban renewal reshaped the landscape.
-
-
Data Collection and Cleaning
-
I worked extensively on gathering and refining historical data, majorly from the Sanborn map of 1951, the Brooklyn Urban Renewal Area Map (20th Century), city directories, and oral histories. This involved cleaning and structuring datasets.
-
-
Multimedia Presentation and Animation
-
I also developed a multimedia presentation to enhance engagement, including an animation of the House of Prayer parade route (View Here). This animation captures the movement and energy of a significant cultural and religious event, helping audiences connect with Brooklyn’s community traditions in an immersive way.
-
The Virtual Black Charlotte project is a groundbreaking effort to digitally reclaim a community that was physically lost but remains deeply rooted in memory. By integrating systemized workflow including mapping, 3D modeling, archival research, and multimedia storytelling, this project creates a dynamic platform for former residents, scholars, and the public to engage with Brooklyn’s rich cultural heritage.
To explore the project, visit Virtual Black Charlotte.
Project name 04: Automating NYC Directory Digitization for Dissertation Research
Researcher
One of the primary sources for my dissertation research is the New York City Directories (1786–1913). These directories provide essential data for tracing the development of the city's art galleries and commercial landscape of the art world. Although the complete collection is available through the New York Public Library (NYPL), it only allows manual, page-by-page downloads, making the process highly time-consuming. I developed an automated solution to efficiently download, process, and prepare these directories for OCR-based text extraction to streamline this.
Automating Data Collection
To systematically download the NYC Directories, I created a Python-based automation script (View Code) that allows users to:
-
Select image quality for downloads.
-
Download entire directories by specifying the range of year.
-
Extract and store images efficiently for further processing.
By leveraging IIIF image IDs from NYPL, the script eliminates the need for manual retrieval, significantly reducing time and effort while ensuring consistent data collection across directories spanning over a century.
Enhancing OCR Readability Through Image Processing
Once all the directories were downloaded, I implemented a preprocessing pipeline to enhance image quality for optimal OCR performance. Using OpenCV and Python, the processing script (View Code) performs:
-
Automated Cropping – Detecting and isolating text regions to remove unnecessary margins.
-
Noise Reduction – Applying denoising filters to improve character recognition accuracy in Tesseract OCR.
-
Batch Processing – Running the pipeline across multiple directories for efficiency.
This step ensures that the directories are clean, optimized, and structured, making text extraction more reliable when running OCR-based data mining.
Impact and Application
By automating data retrieval and preprocessing, this project enables large-scale text analysis of NYC directories, allowing me to trace the spatial and commercial evolution of New York’s art market. This work also sets a foundation for future digital humanities and historical GIS projects, providing an efficient workflow for digitizing and analyzing historical archival records.
Project name 05: Identifying Key Actors in the New York Art Market Through Fuzzy Name Matching
Researcher
For my dissertation research, I developed a Python-based fuzzy matching algorithm to identify key art market actors in New York City Directories (1843–1853) by tracking name variations over time. This method helps determine business stability, revealing individuals and firms that maintained a continuous presence in the art market—a critical indicator of their influence.
Challenges in Historical Data Processing
City directories, while invaluable for tracing commercial activity, present several challenges. Inconsistent name formatting makes it difficult to track businesses across years due to spelling variations, abbreviations, or OCR errors. Additionally, scanned historical documents often contain misread characters or fragmented text, further complicating accurate identification. Manual data cleaning is another major limitation, as reconciling thousands of names across multiple decades is highly time-consuming and prone to human error. To address these issues, I implemented fuzzy name matching using Python, allowing for the automated detection of similar names across different years, even when discrepancies exist in text formatting or OCR outputs.
Project Implementation
The Python script (View Code) utilizes fuzzy string matching techniques to compare names year by year and group similar entries. This method effectively detects spelling variations, abbreviations, and OCR misreads, ensuring accurate linkages between different directory records. By automating name reconciliation, the tool significantly reduces the need for manual data cleaning, streamlining the process of identifying key historical figures in the New York art market. The results from the 1843–1853 dataset (View Results) successfully highlight businesses and individuals who remained active over time, helping to map the long-term presence and influence of art dealers, printmakers, and auctioneers in the city.
Impact and Application
This automated approach saves significant time while ensuring greater accuracy in historical data analysis. By tracking key actors across decades, this research enhances our understanding of business longevity and market stability, shedding light on the economic forces that shaped New York’s evolving art trade. The fuzzy matching tool also holds potential for broader digital humanities applications, particularly in tracking historical figures, businesses, or institutions across inconsistent records. This method provides a scalable solution for longitudinal studies, demonstrating how computational tools can bridge gaps in historical research and archival analysis.
Project name 06: Tracking the Impact of Wikipedia Edits: A Project Vox Initiative
Graduate Assistant
As a part of the team of Project Vox, I contribute to its mission of highlighting philosophical works from historically marginalized voices, particularly those excluded from the European male-centric canon. Project Vox is committed to redefining the history of philosophy by making these thinkers accessible to students, educators, and scholars. One of the ways we achieve this is through Wikipedia Edit-a-thons, where we correct, expand, and refine pages related to historically overlooked philosophers. Given Wikipedia’s global reach, ensuring that these edits remain persistent is crucial to sustaining an accurate and inclusive philosophical record.
Tracking Edit Persistency
I developed a Streamlit-based app in Python to measure the impact of our Wikipedia edits (View Here) which was recently put in action on February 20, 2025, at the Project Vox Symposium to celebrate its 10th anniversary. This app tracks and analyzes the persistence of edits made by users attending the Wiki-Editathon sessions at Project Vox, over time. I aimed to create a dataset that records how long edits by Projrct Vox team, remain visible before being altered or removed. By systematically collecting and processing revision data, this project offers insights into the stability of knowledge contributions on Wikipedia. And it essentially helps to evaluate the effectiveness of Project Vox’s efforts in influencing public discourse.
Methodology
To analyze Wikipedia edits, I built a custom data pipeline that extracts revision histories for 28 pages associated with key philosophers, including:
-
Margaret Cavendish
-
Émilie du Châtelet
-
Anne Conway
-
Juana Inés de la Cruz
For each edit, the tool collects:
-
Editor’s username
-
Timestamp of the edit
-
Time elapsed before the edit is replaced
-
Change in text length (edit delta)
The app automates data retrieval, processes the number of revisions, and calculates persistency metrics, which measure how long edits remain unchanged. Regex-based data cleaning ensures that inconsistencies in usernames and timestamps do not affect the analysis.
Findings
Tracking Wikipedia edits allows us to reckon:
-
Which types of edits persist the longest, and which are frequently overwritten.
-
How different language versions of Wikipedia treat edits related to marginalized philosophers.
-
Who the most active contributors to these pages are.
This information helps Project Vox refine its strategies for ensuring lasting visibility of philosophical contributions, enabling us to advocate for more durable representation in public knowledge spaces.
This project enhances Project Vox’s mission of inclusivity in philosophy by
incorporating data science and automation. The ability to track edit persistency provides valuable insights into Wikipedia’s role in shaping historical narratives and allows us to measure the real-world impact of our work. Through this initiative, we ensure that marginalized voices remain part of the philosophical canon, not just momentarily, but for the long term.
Project name 07: Milestones Across the Color Line (1927)
Graduate Assitant
As part of the exhibition project team for Milestones Across the Color Line (1927) at Duke University Libraries, I collaborated with Professor Victoria Szabo to develop a story map that visualizes and contextualizes Oliver B. Quick’s A Souvenir of Durham, N.C., Showing the Progress of a Race. This project builds upon Mea Warren's prior work and explores the achievements of Durham’s Black community in the early 20th century.
My contribution focused on story map development, integrating historical narratives with interactive spatial visualization to highlight key locations and events. This work enhances public engagement by providing a dynamic, digital experience that showcases Black Durham's rich cultural and historical legacy.
View the exhibition: Milestones Across the Color Line (1927).
Project name 08: The Fuller Company and Interwar Construction (1919–1937)
Graduate Teaching Assistant
As part of the exhibition project team, I collaborated with Susannah Braswell, Oliver Hess, Paul Jaskot, and Elif Ozturk to develop a story map for The Fuller Company and Interwar Construction (1919–1937) (View Here). This project invesigates the legacy of the George A. Fuller Company, a key player in American construction by documenting its prestigious projects between 1882 and 1937, including its early involvement in Duke University's construction.
My primary role in this project focused on data collection and automation, where I designed a workflow that combined traditional research methods with AI-driven data extraction and cleaning. This structured dataset was then used to create the story map visualization, allowing users to dynamically explore the company’s nationwide impact on architectural development.
This project demonstrates how digital tools can enhance architectural history research, making historical construction projects more accessible and interactive for broader audiences.
Project name 09: Durham, Duke, and the World: A Timeline of the 1920s and 1930s
Graduate Teaching Assistant
As part of the exhibition project team, I collaborated with graduate students, an undergraduate, and faculty members from the Art, Art History, and Visual Studies Department to develop an interactive timeline that juxtaposes events at the university, city, national, and global levels during the 1920s and 1930s (View Here). This timeline contextualizes the construction and transformation of Durham and Duke University within the broader architectural, political, technological, and social changes of the period.
My primary contribution was researching and collecting data on key international events from 1918 to 1935 to provide a comprehensive global perspective alongside local and national developments in the timeline. This work helped highlight how international political shifts, technological innovations, and social movements intersected with the evolving urban fabric of Durham and the establishment of Duke University.
This project showcases how visual timelines can effectively illustrate historical interconnections, offering an engaging way for audiences to explore the multilayered history of the early 20th century.