IT, data and delivery · Budapest
Data pipelines, dashboards and the wider projects around them, from the first stakeholder workshop to evaluation.
What I do
Most projects need both: the data has to be right, and people have to use what gets built.
Data
I build pipelines that bring data from source systems into clean, governed tables, then turn it into dashboards, forecasts and answers people can trust.
Delivery
I turn many voices into clear requirements, a realistic plan and a launch people actually adopt, then measure whether it delivered what was promised.
Data · pipeline lifecycle
A pipeline built on Budapest's weather data, following the standard lifecycle: ingestion, storage, validation and transformation, then consumption by reporting (BI), predictions (ML) and AI applications. Swipe or use the arrows to move through the stages.
01 · Ingestion · Collect data from sources
Data arrives in three ways: near real-time feeds every 15 minutes, batch loads of the history back to 1940, and third-party feeds produced by another organisation, here the weather service's forecast.
What I can do
02 · Storage · Store raw data in a lake or warehouse
Each dot is a source row recorded in raw storage before cleaning. Preserve source values alongside ingestion metadata, and keep this layer immutable where possible. This makes transformations reproducible and lets reported values be traced back to their source.
What I can do
03 · Validation · Check completeness, schema and accuracy
Each batch is checked before it is used: completeness (required values are present), schema (the expected fields and types) and validity (values follow agreed rules). Failed rows are set aside with a reason, so issues can be reviewed rather than quietly dropped.
What I can do
04 · Transformation · Clean, join, aggregate and model data
Validated rows are cleaned, joined, aggregated and modelled into tables designed to answer specific questions. Here, the last two complete weeks of hourly readings, up to yesterday, become one row per day: each column shows that day's 24 readings, and the line connects the daily highs.
What I can do
05 · BI · Data consumption · analytics
Each dot shows the highest temperature on 28 October since 1940, grouped in one-degree columns and shaded by decade. The dashed line marks the 1991–2020 normal. The ringed dot tracks today’s forecast high, observed high so far, or final high as the day progresses.
What I can do
06 · ML · Data consumption · predictions
The dots are the daily highs and the line is the weather service's forecast, loaded as a third-party feed. Today's dot stays in the forecast colour until its warmest hour has passed. The rings show what the forecast said, so you can see how far off it was; the band shows how that error grows.
What I can do
07 · AI · Data consumption · AI applications
The same clean, governed tables can power an AI assistant. It retrieves relevant data, answers in plain English and names the tables it used, so people can check the result. Documented definitions, controlled access and evaluation checks help keep answers grounded in trusted information.
What I can do
08 · Governance · Across the whole lifecycle
Public data still needs governance: record its source and licence, assign owners, apply quality rules and control who can change published tables. Every result should be traceable, changes should be reviewed, and users should have only the access they need.
What I can do
Delivery · six phases
A structured approach to project management, from initial planning to final delivery. Adaptable to both agile sprints and traditional Gantt-based plans.
01 · Discover · Stakeholder analysis
Each dot represents a person, team by team. Before anything is designed, I identify who has a stake in the project and what role they play: the sponsor, the process owner, the people who will use the solution, IT, and the supplier. Each is captured in the stakeholder register.
What I can do
02 · Define · Requirements
Each dot represents a requirement, expressed as a user story with clear acceptance criteria. I work with stakeholders to prioritise requirements based on business value, urgency, dependencies, and effort.
What I can do
03 · Design · Process redesign
The expense claim process, mapped as it works today (as-is) and how it could work (to-be). Each dot represents a claim, with clusters showing where work gets held up. Removing duplicate data entry and unnecessary loops for missing receipts can turn weeks into days.
What I can do
04 · Deliver · Planning and control
Each row represents a workstream in the project plan. The line marked Today shows progress against the schedule, filled dots mark completed work, and the diamond marks the go-live milestone. Risks, assumptions, issues, and dependencies are tracked in a RAID log.
What I can do
05 · Adopt · Testing, go-live and change
Before go-live, users test the solution against agreed acceptance criteria through user acceptance testing (UAT), and outstanding issues are resolved before sign-off. After launch, each dot represents a user adopting the new way of working, with the percentage tracking progress.
What I can do
06 · Measure · Benefits realisation
Each bar represents a measure from the business case. The grey mark shows the baseline, the bold mark the target, and the dots the latest results. Two measures have exceeded their targets, while one remains just short.
What I can do
Contact
I work at the intersection of data, BI and business analysis, turning complex needs into practical solutions. If you’d like to discuss an idea, explore a collaboration, or learn more about my work, feel free to get in touch.