DEV Community: Kayvan Zahiri

Your public CSV looks fine until Excel opens it — charset, BOM, and timezone traps

Kayvan Zahiri — Wed, 30 Sep 2026 22:34:35 +0000

Public extracts usually fail in pagination or scope. A quieter killer: the file is “correct” in Python and wrong the moment a buyer opens it.

Before you hand off a public-source CSV, run these five checks:

UTF-8 with an explicit BOM when Excel is in the loop — macOS Numbers / pandas often hide the problem; Windows Excel silently mangling café / São / curly quotes is classic. Prefer utf-8-sig for buyer-facing CSVs, and say so in schema.md.
No smart punctuation from HTML — scrape text often includes NBSP (\u00a0), zero-width chars, and fancy dashes. Normalize whitespace and replace NBSP → space before write.
Timezone is a column, not a vibe — timestamps without an offset become “wrong day” after a buyer’s spreadsheet localizes them. Store ISO-8601 with Z or an explicit offset (e.g. 2026-09-30T15:33:00-07:00), or split into date_utc + time_utc.
Leading zeros survive as text — ZIP codes, FIPS, and external ids. Prefix with a tab or force text quoting so Excel does not turn 02139 into 2139.
Round-trip sample — reopen the shipped file in Excel and LibreOffice once. Spot-check 10 rows for mojibake, shifted dates, and truncated ids. If either client lies, fix the export — not the buyer.

Prefer official JSON/API when it exists; flatten once; document encoding + timezone policy in schema.md next to null rates.

Soft CTA

I ship fixed-scope OpsPacket Public Data Pull packs for public sources only (≤5k rows, ≤12 fields, robots.txt respected) with CSV + schema. Sample shape + landing:

https://kayvan-zahiri.github.io/opspacket-public-data-pull/

Hard no: login-walled SaaS exports, LinkedIn, personal email/phone harvest, ToS-hostile targets. No ranking or traffic promises — just a scoped CSV.

Pagination nulls eat your CSV — a 5-check public extract checklist

Kayvan Zahiri — Wed, 30 Sep 2026 18:57:52 +0000

One-off public extracts usually die in pagination, not in the parser.

Before you ship a CSV from a public HTML/JSON source, run this checklist:

Count pages vs rows — if the site says ~4,200 items and you got 800, you stopped early (or hit a soft wall).
Null spike after page N — plot null rates by page index. A sudden jump usually means the template changed or the bot got a thinner HTML shell.
Stable dedup key — prefer a public id/slug over fuzzy title+city. Re-runs without a key create silent duplicates.
Field cap — keep ≤12 columns. Extra “nice to have” fields are where nested JSON becomes garbage strings.
schema.md — record source pattern, row count, null rates, and what you dropped. Future-you (or a buyer) will thank you.

Prefer official JSON/API when it exists. Flatten nested fields once and note the drop.

Soft offer

I sell a fixed-scope OpsPacket Public Data Pull for public sources only:

≤5,000 rows · ≤12 fields · robots.txt respected
CSV + schema.md
$149 ≤72h · $199 rush ≤24h

Landing + sample CSV: https://kayvan-zahiri.github.io/opspacket-public-data-pull/

Checkout: https://kayvanandre.gumroad.com/l/public-data-pull · rush https://kayvanandre.gumroad.com/l/public-data-pull-rush

Contact: kayvanandre@gmail.com

Hard no: login walls, CAPTCHA farms, LinkedIn, personal email/phone harvest, ToS-hostile targets.

Before you build a scraper: a public-data CSV scope card that keeps one-offs honest

Kayvan Zahiri — Mon, 28 Sep 2026 17:47:57 +0000

Most “quick scrapes” fail the same way: unclear fields, no row cap, and silent nulls after pagination.

Before you open Playwright (or hire someone), write this scope card for any public source:

URL pattern — exact pages or API endpoints (include pagination / sitemap).
Public-only check — no login wall, no CAPTCHA farm, robots.txt / ToS allow automated fetch.
Fields (≤12) — name each column; drop “nice to have” extras.
Row budget — e.g. ≤5,000 rows. Caps force honesty about coverage.
Dedup key — what makes two rows the same (id, url, …).
Delivery — CSV plus a one-page schema.md with row count, null rates, and the source pattern.

Prefer official JSON/API when it exists (?$limit= on open portals often beats HTML parse). Flatten nested fields once; note what you dropped.

Sample shape

I keep a public Open Brewery DB (California) slice as an example CSV + schema on the landing below so buyers (and future-me) can see the output shape before paying.

Soft offer

I’m the maker of OpsPacket Public Data Pull — fixed-scope public extracts only:

≤5,000 rows · ≤12 fields · robots.txt respected
CSV + schema.md
$149 within 72 hours · $199 rush within 24 hours

Landing + sample: https://kayvan-zahiri.github.io/opspacket-public-data-pull/

Checkout: https://kayvanandre.gumroad.com/l/public-data-pull · rush https://kayvanandre.gumroad.com/l/public-data-pull-rush

Hard no: login-walled SaaS exports, LinkedIn, personal email/phone harvest lists, ToS-hostile targets. Paste a public URL + fields and I’ll say fit / out-of-scope in one reply.

No ranking, traffic, or revenue guarantees — just a scoped CSV.

A homepage SEO punch-list that doesn’t become a three-week rabbit hole

Kayvan Zahiri — Thu, 24 Sep 2026 21:53:03 +0000

Most SEO audits dump hundreds of warnings. Owners still don’t know what to fix first.

Before you open Screaming Frog or buy a tool, run this public-HTML pass on the homepage (no CMS login):

Title — present? ~50–60 characters? Not “Home”?
Meta description — present? Roughly under ~155 characters?
H1 — exactly one, matching what the page sells?
Image alts — content images missing empty alts?
Basics — canonical; accidental noindex; status 200

Sort into P0 / P1 / P2. Ignore vanity scores until those are clean.

Self-serve checklist: https://kayvan-zahiri.github.io/opspacket-seo-snapshot/self-serve-punch-list.md

I’m the maker of OpsPacket. If you want that pass done for you on a public URL you authorize, there’s a free punch-list form: https://opspacket.com/#free-sample

Optional paid $49 Site Health Snapshot (up to 15 same-site URLs, status CSV, capped top-5 fixes): https://kayvanandre.gumroad.com/l/seo-snapshot

No ranking, traffic, or revenue guarantees.

75% of resumes get rejected by ATS before a human sees them — here's how to check yours for

Kayvan Zahiri — Sat, 14 Mar 2026 22:51:07 +0000

Hey everyone, I've been frustrated with how many qualified people get rejected by ATS before a human ever
sees their resume. So I built a free tool that scores your resume against any job description.

How it works:

Paste your resume text + the job description
Get a score out of 100 across 5 categories (keyword match, formatting, sections, content quality, length)
See exactly which keywords you're missing
Get specific suggestions to improve

It's completely free, no sign-up required: https://withresumeai.com/ats-checker

There's also a free AI resume builder if you want to start from scratch:
https://withresumeai.com/builder/new

Would love feedback on what could make it more useful!