Clean once
Every source is mapped to one schema, so metrics mean the same thing everywhere.
All work
Concept project: a sample build that shows how we approach this kind of product. The client, features and figures are illustrative.
Full stack · Python (FastAPI) + Next.js · Analytics · Retail
insightiq
A real-time BI dashboard that pulls sales data from every channel into one consistent set of numbers.
Overview
The brief
A retailer selling through its website, two marketplaces, shops and wholesale had a different revenue number in every report. Each Monday an analyst spent the morning exporting CSVs, cleaning them in spreadsheets and pasting charts into a deck that was out of date by Wednesday. InsightIQ collects the data automatically, cleans it once, and gives every team the same live dashboards.
Problems & fixes
What was in the way
Pick a problem to see what was wrong, what we built and what changed.
The problem
Shopify, Stripe, marketplace exports and shop sheets used different product names, currencies and refund rules, so totals never matched.
What we built
Pandas ETL jobs that map every source to one schema: product and channel mapping, currency conversion, refunds netted the same way, validated with Pydantic before loading.
What changed
One agreed definition for revenue, orders and margin across every report.
1 / 4
Impact
What it changed
5Data sources unified
8 hrsReporting time saved weekly
<600 msTypical dashboard load
1Agreed revenue number
- Leadership, sales and finance look at the same numbers.
- Reports refresh themselves instead of being rebuilt every Monday.
- Problems like a falling repeat rate are seen in days, not at quarter end.
How it works
Under the hood
Click any part of the system, or play the flow from start to finish.
Step 1 · APIs · CSV upload
Connectors
Scheduled pulls from Shopify, Stripe and marketplaces, plus validated CSV uploads stored in S3.
Drill-down
From a monthly total down to the channel, product and order behind it.
Export-ready
Any board exports to CSV or a branded PDF for board meetings.
Timeline
- Week 1Metric definitions agreed with each team, source audit
- Weeks 2–3Connectors, ETL jobs and data validation
- Weeks 4–5FastAPI layer, caching and Next.js dashboards
- Week 6Permissions, exports, AWS deployment and monitoring
Tech stack
PythonFastAPIPydanticSQLAlchemyPandasNext.jsTypeScriptRechartsPostgreSQLRedisDockerAWS
Tell us what you're building. A senior engineer, not a salesperson, replies within one business day.
Start a project