Vol. I … No. 1Butwal, Nepal — Friday, October 30, 2026 — Evening EditionPrice: One Conversation
Notice
Now accepting remote data-engineering internships for 2026–27.
The Daily Abiral
On Github
Fifteen repositories, all of them his own work. Read the archive →
“All the data that’s fit to pipeline.”
Stop Press•ghtrends reads GitHub hourly — stars & pull requests•Ledger-Sentinel keeps watch over every entry•Job-board scraper joins the fleet, tests included•296 contributions in the past year•Open to remote data-engineering internships•Stop Press•ghtrends reads GitHub hourly — stars & pull requests•Ledger-Sentinel keeps watch over every entry•Job-board scraper joins the fleet, tests included•296 contributions in the past year•Open to remote data-engineering internships•
The Lead Story — Data Engineering
I Take Raw Data at One End and Hand Back Answers at the Other
I’m Abiral Pokhrel from Butwal, Nepal. I build data systems in public — and I finish what I start.
I builddata pipelines that run on schedule
The Record
Contributions, past year
296
Repositories
15
Public projects
11
Private projects
4
Degrees in progress
1
Group members remaining
1
Figures from the GitHub record, surveyed October 2026. The last line is explained in the story at right; he did the work anyway.
The works, as our engraver imagines them: raw events enter at the funnel, are sifted, stored in columns, and leave as tables a person can actually read. Engraving for The Daily Abiral
By Abiral Pokhrel
Staff correspondent — and the entire staff
Butwal, Nepal — I’m Abiral. I study Computer Systems Engineering at ISMT College Butwal (University of Sunderland) — graduating June 2028. My real classroom is GitHub: everything I build is open to inspect.
Messy data in, clean answers out. That’s the work I love. When my group left our enterprise project, I finished it solo. I ship, I document, I move on to the next build.
Now: a remote data engineering internship. Bring me a real pipeline — I’ll bring the hands that finish it.
It collects listings, cleans them, and pulls from the Remotive API plus a dynamic browser collector. Three tidy commits, tests included, data kept local.
Listings in, tested data out. Animated plate — Scraper
No more hopping around the headlines — he has a proper job now. Watch him work the way the work actually goes: sweep the mess out first, carry the data across, then get the wrench on the pipeline. Each scene runs its little story on a loop. Tap a scene to pause or resume it.
Job No. 1 — Cleaningmessy rows ✕clean ✓
1 · Find the mess2 · Sweep it out3 · Floor’s clean
Sweeping the floorCleaning the Data
He works the floor in passes — dust up, messy rows out — and only stops when the clean tag pops. Duplicates, blanks and broken rows get swept out first; nothing gets modelled until the floor is clean.
Job No. 2 — Transfer
Raw Data
Clean Answers
1 · Pick it up2 · Carry it across3 · Delivered
Jumping title to titleMoving the Data
Crouch, launch, land — platform to platform, and back again for the next batch. The packets fly the same arc he does: source to storage to model to answer, stage by stage.
Job No. 3 — Buildingleaking join ✕
✓
1 · Spot the leak2 · Wrench it3 · Sealed
Making & fixingFixing Data, Building Pipelines
Three good taps on the joint — sparks, a shudder, gears meshing — and the drip stops, the flow runs, the seal pops. When a pipeline breaks, this is the job: fix it, test it, ship it.
One mascot, three jobs — sweeping, carrying, fixing · Tap a scene to pause it
The Projects
Page A2 — Works of Record
Make it yours
Reading spacing
Show:
Showing all three stories
Serverless — AWS
ghtrends: The World’s Code, Read Hourly
A serverless pipeline that listens to GitHub around the clock and reports back on what the world is building.
Hourly ingestServerless AWSTop 500 searchable
Every hour it ingests the GitHub Archive — stars and pull requests — and lands Parquet in S3.
Glue catalogues it. dbt models it into a daily trends table. Streamlit shows it — searchable by meaning, not just by name. No server stays awake all night.
Tap a stage — see what it does
GH Archive — the hourly firehose: every star and pull request, captured as it happens.
The works in motion: a pulse rides the wire as each stage lights in turn — hourly in, searchable out. Animated plate — ghtrends
Ledgers fail quietly — a row entered twice, a figure out of place. This sentinel reads every entry, runs its checks, and flags the trouble instead of letting it pass.
Every entry checkedDuplicates flaggedAll checks reported
Each entry is watched as it lands: balances reconciled, gaps hunted down, amounts kept in range.
When the same invoice turns up twice, it doesn’t slip through — it gets flagged, in the open, while the rest of the book passes its checks.
The watch, in motion: entries scanned one by one, the duplicate caught and flagged, the rest passed. Animated plate — ledger-sentinel
Corrections — An earlier edition of this newspaper’s subject overstated the appetite of ghtrends, claiming it ingested every GitHub event. It ingests stars and pull requests, hourly, and quite enough of them. The record is corrected with pleasure.
Memes I Save for No Reason
Page A7 — The Back Page
The Daily AbiralThe Back Page — Memes
Memes I Save for No Reason
Fifteen exhibits from the editor’s phone, entered into the record without explanation.
