asabisoft

Data science/Analytics engineering/Commerce

Data science that leaves the notebook.

Asabisoft is a data science and analytics consultancy. We build LTV, churn and demand models, the pipelines that feed them, and the dashboards and products that put them to work — led by a data scientist with 13+ years in analytics, engineering and product.

Book a 30-minute call See the track record
Illustrative chart: observed revenue per customer for twelve months, a fitted curve, and a forecast with an 80% interval for the following twelve months.02040608003691215182124observedtodayforecast · 80% interval
Fig. 0 Cumulative revenue per customer, months since acquisition — the kind of model we put in production.Illustrative data · hover or drag to read
  • Python
  • SQL
  • LTV models
  • Churn
  • BigQuery
  • Snowflake
  • PostgreSQL
  • A/B tests
  • Forecasting
  • Power BI
  • Tableau
  • Celery
  • Playwright
  • Shopify
  • Stripe
  • LLM automation
  • Wasted ad spend
13+ yrs
in data science, analytics and engineering
1.25M+
installs on products steered by analytics
8+
web and mobile products launched
1M+
URLs a day through our own data pipeline
Fig. 01 — Services

Models first. Then the plumbing and the product around them.

Hire one practice, or all three so the data, the models and the software share the same definitions.

01 / predictsurvival · forecast

Data science

Models that answer money questions: what a customer is worth, who is about to leave, what to launch and what happens if the price changes.

  • LTV, churn and demand forecasting
  • Experiment design and A/B test analysis
  • Economy, pricing and scenario simulation
  • Launch / green-light decision models
  • ML and LLM automation in production
02 / measureevents → warehouse

Analytics engineering

One trustworthy set of numbers: event instrumentation, pipelines, data quality and dashboards your team actually opens.

  • Event tracking design across product and web
  • Pipelines into BigQuery, Snowflake or Postgres
  • Data validators and automated quality tests
  • KPI and unit-economics dashboards
  • Scraping, market and GIS data collection
03 / shipfunnel · store · saas

Commerce & product engineering

The software around the data: Shopify stores and apps, SaaS products and internal tools built in Python.

  • Shopify builds, apps and migrations
  • Store speed and conversion work
  • SaaS and internal tools (Flask, Celery)
  • Crawling and automation with Playwright
  • Stripe billing and integrations
Fig. 02 — Track record

Models that made real decisions.

The question, the model built to answer it, and where it ran.

#QuestionModel & methodWhere
01Which games deserve a launch budget?Green-light / red-light models on early KPIs, with LTV, CPI and ROI pipelines feeding publishing decisions.N3TWORK Studios
02How will a token economy behave before it ships?Simulation models of user-economy behaviour and revenue impact, on top of event instrumentation for wallets, bridges and exchanges.Forte
03Which products and channels actually pay back?Automated LTV, ROI, EBITDA and churn forecasting over terabytes of product data, steering a multi-million-dollar ad budget.Jawabkom
04Is the data even right?Data validators, quality standards and automated test pipelines across every product team’s metrics.Forte
05How healthy is this online store?Store health scoring from crawls, Lighthouse and validation data, served as real-time KPI dashboards.URLAudit
06How much should we produce, and when?Data-driven production planning, inventory and sales forecasting for an agricultural producer.Sultan Sera
Fig. 03 — Selected work

Products we built and still run in production.

SaaS, internal tools and client builds — most of them live on our own infrastructure today.

SaaSFounder

URLAudit

Automated e-commerce store auditing. A distributed Playwright crawler checks 1M+ URLs a day for HTML/CSS errors, Lighthouse scores and visual regressions, with usage-based Stripe billing.

  • Python
  • Celery
  • Playwright
  • PostgreSQL
AI toolBuilt

Creative Generator

Give it a Shopify URL; it scrapes products, extracts the brand identity and produces platform-ready ad creatives and short videos with multi-model planning.

  • Shopify
  • Gemini
  • Veo
  • Claude
ManufacturingBuilt

MAS ERP

Production management for factories: visual floor plans, source-to-product traceability and multi-level BOMs — first deployed in dairy processing.

  • ERP
  • Traceability
  • Dashboards
E-commerce appClient

November Jewel

A live engraving editor for a jewellery store on ikas: customers preview their text on the piece before ordering.

  • ikas
  • Product customiser
Open sourceMaintainer

Master Dashboard

A canvas for running many AI coding agents, terminals and browsers side by side — the tool this studio runs on. Published on npm.

  • TypeScript
  • Agents
  • npm
Data productsBuilt

Scraping & GIS

Scrapers and visualisations for housing-market GIS maps, stock markets, classifieds and gaming data.

  • Scrapy
  • Pandas
  • MongoDB
  • QGIS
Fig. 04 — Process

Small, fixed-scope steps. Stop after any of them.

No retainers by default. Each step ends with something you own.

  1. 01 — week 1

    Diagnose

    Your data, tracking and existing models reviewed end to end. You get a ranked list of what to fix and what to build.

  2. 02 — weeks 2–6

    Build

    Pipelines, models and dashboards shipped in small, reviewable pieces — in your stack and your accounts.

  3. 03 — ongoing

    Measure

    Every model and change gets a metric and a test, so you can see whether it earns its keep.

  4. 04 — anytime

    Hand over

    Documented code, owned infrastructure and a walkthrough, so your team can run it without us.

Portrait of Arif Sait Birincioglu, founder of AsabisoftARIF SAIT BIRINCIOGLU
founder · data scientist
Fig. 05 — the human in the loopIzmir · remote
Fig. 05 — About

You work with the person who writes the models and ships the code.

I’m Arif — data scientist, Python engineer and founder of URLAudit. Before Asabisoft I led analytics for blockchain games at Forte, built LTV and green-light models at N3TWORK Studios, and ran product, data and a multi-million-dollar ad budget at Jawabkom, where we launched 8+ web and mobile products.

I hold an MSc in Systems Engineering and have spent my career at the point where models meet money: launch decisions, ad spend, pricing and production plans. Asabisoft offers that as a service, with the engineering to put the models into production.

Forte
Data analytics lead, games & blockchain
N3TWORK Studios
Senior data scientist, platform insights
Jawabkom
Data scientist & senior program manager
MSc Systems Engineering
Gediz University · TUBITAK grant
Fig. 06 — Writing

Field notes from real projects.

All articles
Data science10 minLTV forecasting with under a year of dataWhich customer lifetime value models hold up on short purchase histories, how to sanity-check them, and when a cohort table beats all of them.Commerce8 minWhat we check before touching a Shopify themeA redesign fixes how a store looks, rarely why it underperforms. The audit order we run first — speed, tracking, funnel and unit economics — and why each comes before the next.
Fig. 07 — FAQ

Straight answers.

The questions people ask on the first call.

What does Asabisoft do?

Asabisoft is a data science and analytics consultancy. It builds predictive models such as LTV, churn and demand forecasts, the data pipelines and dashboards behind them, and the Shopify stores, SaaS products and tools that use them.

Who will I actually work with?

Arif Sait Birincioglu, the founder — a data scientist and Python engineer with 13+ years across analytics, engineering and product, including Forte, N3TWORK Studios and Jawabkom.

Where are you based?

Izmir, Turkey. Work is remote, with clients in the US, Europe and Australia, and overlap with both European and US Eastern hours.

How does an engagement start?

With a short, fixed-scope diagnosis of your data, tracking or models. It ends with a ranked list of fixes; you can take it in-house or have us build it.

Which tools do you use?

Python, SQL, PostgreSQL, BigQuery, Snowflake, MongoDB, Redis and Celery for data; Power BI and Tableau for reporting; Shopify, Flask, Playwright and Stripe for products; and LLM APIs where they help.

Tell us what the numbers should say.

A 30-minute call, no deck. You leave with a short list of what to model or fix first — whether or not we work together.

Book a call Message on LinkedIn