Service Format
Problem
Data is spread across files with duplicates and manual handling, so reporting stays slow and unreliable.
Solution
I unify sources into one pipeline: cleaning, normalization, enrichment, and export to your target format.
Result
A ready dataset for CRM, analytics, or campaigns without daily manual spreadsheet assembly.
Timeline
After the brief and a sample file, we fix phases: pilot on a subset, then full run on total volume.
Collaboration Format
You send the brief and sample data → we align processing rules → I show a test output → we run full volume.
Start with the brief
How It Works in Practice
When data is spread across many files, business decisions become guesswork. I help you build one trusted data flow.
When This Is Relevant
- Reporting is manual and frequently inconsistent with real operations.
- Exports contain duplicates, noise, and conflicting formats.
- Your team needs faster preparation for CRM, analytics, and marketing.
What I Do in the Project
- I unify fragmented sources into a rule-based processing flow.
- I implement cleaning, normalization, and enrichment with clear logic.
- I deliver target-format exports without post-processing pain.
What You Receive
- A clean and structured dataset ready for business use.
- A repeatable processing pipeline that scales with volume.
- Transparent transformation rules for every output field.
We usually start with a test sample to validate the approach quickly, then apply the same logic to full volume.
Service Materials and Case Studies
May 31, 2026
A delivered project write-up: work on the software layer of an urban air monitoring system in Saint Petersburg, covering monitoring stations, gas analyzers, Tiny Core Linux, documentation, and controlled code delivery.
Read material
May 30, 2026
A delivered project write-up: moving an old Perl application from a single-server setup to UpCloud, migrating MySQL, preparing object storage for images, and managing the cutover point carefully.
Read material
May 29, 2026
A case study on preparing an English LaTeX version of R. A. Fisher's classic iris paper: document structure, formulas, tables, TikZ graphics, rendered pages, and raw TeX source.
Read material
May 26, 2026
A case study about a municipal traffic accident map for Untolovo: turning traffic police data for 2016-2020 into a clear map, statistics on pedestrians, cyclists, and children, and a tool for public oversight of road ...
Read material
May 25, 2026
A case study about a city map of smoke-free cafes and restaurants in Novi Sad: turning scattered local knowledge into a structured directory with filters, a map, a Viber bot for feedback, and a practical user flow.
Read material
May 20, 2026
A case study of the Viber bot for Smoke Free Cafes Novi Sad: language and category selection, geolocation request, 7 nearest smoke-free venues, and a direct channel to project administrators.
Read material
April 4, 2026
How order picking verification was automated: an employee sends invoice and product photos, OpenAI compares them and returns a structured JSON conclusion, and the Telegram bot returns the result per item.
Read material
April 3, 2026
A practical PoC of fully automated EN->RU translation with Qwen 2.5 32B on Mac: CPU mode was too slow, GPU mode was production-viable, with quality control for problematic chunks.
Read material
April 2, 2026
A practical DeepSeek OCR 2 test on a maritime charter book: many headings, links, and footnotes, with final assembly into a single LaTeX file.
Read material
February 23, 2026
I rebuilt an email sync process from an external API in n8n: removed conflicting parallel runs, added export-readiness waiting and API rate-limit handling, and merged load + cleanup into one stable flow.
Read material
February 19, 2026
We converted raw real estate deals into a practical Supabase schema, linked projects and developers, added time analytics fields, sped up queries with indexes, and prepared a flat mart for Excel and external API use.
Read material
January 6, 2026
I automated translation of 3,000 route names and descriptions from Excel into Russian using n8n and OpenAI, with intermediate quality checks, preserved file structure, and staged execution (test, 500 rows, full run).
Read material
January 2, 2026
A case study about building a list of city districts across Russia: I first expected a single SQL query against "КЛАДР", then tried "ГАР" (a ~50 GB archive with download and unpacking constraints), and ultimately deli...
Read material
January 27, 2021
Practical notes on Unicode and UTF-8 in Perl: core concepts, encodings, Encode APIs, and common pitfalls when processing text.
Read material
June 10, 2016
GPA Calculator: a small web tool for converting course grades and credits into one weighted average indicator.
Read material
May 29, 2016
Story of building a weather archive for Patagonia summits: parsing mountain-forecast, storing daily data in CSV, and publishing chart views for temperature, wind, and snow.
Read material
April 20, 2015
Fibus: an aggregator idea between private carriers and passengers who preferred minibuses over regular buses for trips between Saint Petersburg and Finland.
Read material
October 4, 2014
A practical example from 2014: moving a complex ceiling-sales pricing workflow from manual calculations/Excel to a browser calculator with instant recalculation, hints, and fewer errors.
Read material
September 15, 2014
A postal delivery calculator from Saint Petersburg: a calculation API for an online store, with a web form built on top of the same service.
Read material
August 20, 2014
Donorium: a blood donor database that grew from a Leprosorium.ru post into an open service for finding donors and sending blood donation requests.
Read material
June 10, 2014
A web tool for filling a complex Russian Post 112-EP PDF form: users entered data in a simple interface, received a completed document, and could print it immediately.
Read material
May 25, 2014
A practical case study about processing several hundred thousand e-commerce images: watermark generation, resizing, AWS S3 storage, and a custom Perl + ImageMagick pipeline from the pre-Cloudflare Images era.
Read material
Discuss a similar task