Daniel MarchanSoftware Engineer

Demo Video

Tech stack

• Client: React 19, TanStack Router, TanStack Query, Zustand, Base UI, Tailwind CSS v4, MapLibre, Vite
• Server: NestJS 10, Express, nestjs-trpc, Drizzle ORM, Better Auth, Zod, neverthrow
• Data: PostgreSQL, Redis + BullMQ
• Infra: pnpm workspaces, Turborepo, Docker, Vitest

Design decisions

• Vendor data is an evictable cache, not a system of record. Listing rows are shared between users and swept on a retention schedule, so saving a property copies a frozen snapshot rather than foreign-keying into the cache. Cache eviction can never touch user-owned data.
• Underwriting assumptions freeze at save. A saved property captures the user's financing and operating assumptions at the moment of saving, so its cash flow stays a stable artifact instead of silently re-pricing every time the search filters change.
• Enrichment progress streams over SSE. Rent estimation is slow and per-property, so metrics push to the client through a tRPC subscription as they resolve rather than blocking the initial map render.
• Every vendor sits behind a provider boundary. Persistence stays vendor-neutral so the data source can be swapped without touching domain logic.

Demo Video

Tech stack

• Client: React 19, React Router 7, TanStack Query, Tailwind CSS v4, Vite
• Server: Express 5, Drizzle ORM, Better Auth, Zod
• Database: PostgreSQL 16
• Infra: pnpm workspaces, Turborepo, Docker, Fly.io

Try the demo

Open www.datadrop-app.com and sign in with one of the accounts below. Both demo users share the same password: DataDropPass123.

Admin — [email protected]

Signs in with full access to the admin surface. Things to try:

• Go to Users to browse, create, and edit users (admin or viewer).
• Go to Datasets to see the seeded eCommerce Store Performance dataset, then open its Uploads page to upload a CSV and assign the dataset to viewer users.
• Switch over to the viewer surface to see how the data looks to the people you just granted access to.

Sample CSVs

Download any of these to upload from the admin Uploads page:

• orders.csv — baseline valid dataset
• orders-variant-1.csv
• orders-variant-2.csv
• orders-variant-3.csv
• orders-variant-4.csv
• orders-variant-5.csv
• orders-invalid.csv — intentionally malformed, to see validation errors

Viewer — [email protected]

Signs in with read-only access to datasets assigned to them. Things to try:

• Go to Datasets to see the datasets the admin has shared.
• Open a dataset to browse rows with pagination.
• Open the Charts view for a dataset to visualize uploaded data.