Context: Our Directory Workflow (Read before
answering)
We are building a community-verified directory (Practo-for-Animals style) for
categories like vets, shelters/NGOs, feeders, boarding, trainers, groomers, pet
shops, etc. Many of these are unorganised/local businesses, so information will
often be incomplete or inconsistent.
How data comes in (Primary Layer)
Listings will be sourced through:
Intern research (Google Maps, Instagram pages, Justdial, Facebook
groups, websites, etc.)
Community submissions (WhatsApp, forms, DMs, referrals)
Existing Pawzz network inputs (rescuers, feeders, partners)
You scraping the data yourself
Q1) Tools + System Design for Community-Verified
Directory (Long Answer)
We are building a community-verified directory of unorganised/local businesses
(vets, shelters/NGOs, feeders, boarding, trainers, groomers, etc.). Data will
come from interns + public sources (google, social media scraping) +
community submissions, so it will be messy and inconsistent.
Question:
If you were the lead Data Analyst for this directory, how would you design the
entire data collection + cleaning + verification workflow so that listings are
accurate, scalable, and auditable?
In your answer, explain which tools you would use and why (Google
Sheets/Excel, SQL, Python, Power BI, Airtable/CRM, forms, automations, etc.)
and how you would structure:
required fields & proof standards per category
data pipeline stages (submitted → needs fix → ready → verified →
published)
dashboards/KPIs to track throughput, quality, and backlog
how you would handle unstructured inputs (WhatsApp messages,
screenshots, inconsistent addresses/phone formats)
(Answer in a detailed plan — assume you are building it for a real team and
must reduce founder dependency.)
Q2) Red Flags + Verification + Duplicate Prevention
(Long Answer)
Because listings are from unorganised/local sources, we’ll face
fake/incorrect data, duplicates, and inconsistent details.
Question:
What are the top red flags you would use to detect unreliable listings, and what
would your verification + duplicate-prevention system look like end-to-end?
In your answer, cover:
your verification checklist for at least one category (e.g., vet clinic or
NGO)
how you’d handle conflicts (two different phone numbers/addresses for
same business)
how you’d design an audit trail (who verified, when, what proof, what
changed)
escalation rules: when to reject, when to re-check, and when to publish
with warnings