LoomHer: research, design, and digital quality review
How could women-led textile production in Uganda become reliable enough for buyers and credible enough for funders? My NYU capstone examined this question through desk research, eight semi-structured practitioner interviews, and applied venture design.
Explore the interactive research prototype
What changed through the research
The original concept centered on portable looms and production at home. The interviews raised questions about buyer demand, quality consistency, workflow coordination, and financing equipment before dependable orders existed.
I revised the proposed first phase toward a managed workshop with shared looms, production supervision, documented quality review, and a limited product range. This change made production coordination and buyer requirements central to the model.
What I produced
The project produced a capstone report, an operating-model proposal, and a browser-based proof of concept for comparing textile images and recording review results. The prototype explores how digital records could support human review and traceability.
Current status and limits
LoomHer is a completed academic project and proposed enterprise model. It is not an operating textile business. The prototype has not been independently validated as a certification system or commercial quality-control product. It does not establish labor conditions, ethical sourcing, material composition, durability, or producer income.
Practitioner conversations informed the design. Participation does not imply endorsement of the final proposal.
LoomHer Quality & Traceability
Interactive Research Prototype
The LoomHer Quality & Traceability Interactive Research Prototype explores how design references, computer-vision-assisted comparison, human judgment, and QR-linked records can support textile quality documentation and traceability. It extends the LoomHer Empowerment project developed through my NYU capstone research.
The prototype makes human oversight visible: image comparison informs a review, a person records a reasoned decision, and publication is a separate choice. It connects my interests in responsible AI, cyber ethics, and applied policy research with questions of craft and women’s economic participation.
What you can explore
- Create and save textile design references with a reference image and intended specifications.
- Link finished-textile photo submissions to their design references.
- Compare color palettes and visual structure, with image-quality cautions to inform the review.
- Keep computer-vision results separate from the human decision. Scores are comparison indicators, not probabilities, objective quality grades, or certified measurements.
- Record a reasoned human decision: approval for demonstration, a rework request, rejection, or a request for a better photograph.
- After human approval, explicitly publish the listed product information as a QR-linked public demonstration record.
- Inspect a demonstration audit history of submissions, comparisons, decisions, publication, and withdrawal. Erasing a workspace removes its data and audit history.
- Try synthetic LoomHer sample images and records to explore the workflow without uploading your own material.
How It Works
Start with the public demonstration workspace and its synthetic samples, or add material you have permission to use.
Create a design reference
Add the intended colors, dimensions, material declaration, pattern notes, and reference image.
Submit a finished textile
Link the textile to its design reference and upload a photograph.
Run the visual comparison
Compare color palettes and visual structure and review any image-quality cautions.
Conduct human review
Inspect the images, consider the method’s limitations, and record a reasoned decision.
Publish a demonstration record
After human approval, review the listed public fields and choose whether to publish the permitted product information.
Share and trace
Use the generated QR code or public link to open the demonstration traceability record.
Research boundaries
- Sample textile images and records are synthetic and illustrative. The prototype has not been validated through real-world field testing.
- It does not verify fiber content, dimensions, durability, labor conditions, geographic origin, or regulatory conformity. Entered specifications are declarations, not verified measurements.
- A QR code identifies a demonstration record; it is not an authenticity guarantee. Public records can be withdrawn.
- A production deployment would require field testing, validated capture procedures, verified user roles, security review, and an appropriate data-governance process.