I help teams replace manual reporting, scattered workflows, and unreliable data processes with automated, documented systems.
Illustrative example
I'm Jiyoung Roh, a Data Systems & Automation Consultant. I can sit with stakeholders, understand where reporting or workflow is breaking down, and then build the Python/SQL pipeline, automation workflow, or reporting system that fixes it.
Most of my work starts the same way: a team has outgrown its spreadsheets, reporting is fragmented, or someone is spending 10 hours a week on something that should take 10 minutes. I help clarify whether the right solution is a dashboard, an automation workflow, a data pipeline, or simply a cleaner process — then build it.
I understand the business problem and can implement the technical solution — no translation layer needed.
Everything I build comes with documentation. Your team should be able to maintain it after I leave.
Available via Upwork, direct freelance, or consulting engagements.
Before building anything, I map where time is going — manual reporting, spreadsheet cleanup, broken handoffs. You get a clear picture of what to fix first.
PDF extraction, spreadsheet cleanup, scheduled queries, data validation — if someone on your team is doing it by hand every week, I automate it.
Data needs to move from source to warehouse to dashboard without breaking or losing rows. I build the pipelines, set up the warehouse, and structure the reporting layer.
Reports live in 5 different places and nobody trusts the numbers. I build a single reporting layer your team can rely on — Looker Studio, Metabase, or the right tool for the job.
Legacy system migrations, platform transitions, or project handoffs — I handle the data cleaning, migration logic, validation, and documented handoff so your team can maintain it after I leave.
A Canadian fitness chain with 50+ locations had no unified reporting. I built a BigQuery reporting layer, set up Apps Script refreshes to keep data current, and delivered Looker Studio dashboards for cross-location visibility.
A U.S. insurance firm was manually keying data from hundreds of claim PDFs weekly — multiple formats, inconsistent layouts. I built Python extraction with regex/parsing logic to handle each format, then piped structured output into Excel VBA for downstream use.
A U.S.-based workforce intelligence platform needed to go from raw, high-volume datasets to self-serve analytics. I built a DuckDB + Parquet processing workflow, curated reporting layers, and stood up Metabase dashboards for self-serve BI.
This project involved moving large volumes of data into a more reliable structure while minimizing disruption to ongoing operations. Data consistency and migration stability were both critical.
Real numbers from real projects — not hypothetical benchmarks.
BigQuery + Apps Script + Looker Studio replaced 10+ hrs/week of manual work at a 50-location Canadian fitness chain
DuckDB + Parquet + Metabase workflow built for a U.S.-based workforce intelligence platform — structured large-scale datasets into self-serve reporting layers
HYBE data migration — automated workflow with validation steps ensured data consistency and minimal operational disruption
Python + regex parsing replaced manual PDF data entry for an insurance client — multiple formats handled, Excel VBA output for the ops team
Tell me what's breaking — the reports, the spreadsheets, the PDFs, or the workflow. I'll tell you honestly whether I can help and what the approach would look like.
Tell me what's breakingLoom walkthroughs instead of hour-long meetings. Written updates, not status calls. I've worked across KST, EST, and PST — timezone flexibility is built in.
I understand the reporting pain and build the technical fix — no middleman needed. You describe the problem in business terms, I deliver the pipeline, dashboard, or automation.
Everything I build comes with documentation your team can follow. The goal is that you don't need me after the project ends — the system runs on its own.