QSpark

Reliable data. Clear decisions.

I help small and medium-sized businesses connect, process and make use of their data.

I specialise in Databricks and Apache Spark on Microsoft Azure.

I build data solutions, Power BI reports and automations that cut down everyday manual work.

Services

How I can help

From getting hold of your data, through processing it, to reports and automation. Every solution is designed around the systems you already use and what you need to achieve.

  • Data extraction and integration

    The problem
    Your data is scattered across databases, applications, APIs and files.
    What I deliver
    I design and set up regular loading of data into one shared environment.
    The benefit
    Less manual exporting, and data that is ready for further processing.
  • Data processing and automated pipelines

    The problem
    Data is cleaned and combined by hand, or processing is unreliable.
    What I deliver
    Automated transformations in Apache Spark and Databricks, with scheduled runs and recovery from failures.
    The benefit
    Repeatable processing and data outputs that no longer need manual work.
  • Data storage and lakehouse design

    The problem
    Your data keeps growing, but has no clear structure or single place where it lives.
    What I deliver
    Design of layers, data models and storage on Azure and Databricks.
    The benefit
    A well-organised data platform to build analytics and future work on.
  • Data preparation and models for reporting

    The problem
    Different departments calculate the same figures in different ways.
    What I deliver
    Analytical data models, shared metric definitions and ready-to-use datasets.
    The benefit
    A consistent foundation for reports and easier evaluation.
  • Power BI reports and dashboards

    The problem
    Key numbers are buried in countless spreadsheets and files.
    What I deliver
    Report design, data model, metrics and automatic data refresh.
    The benefit
    An overview of how your business is doing, all in one place.
  • Performance and cloud cost optimisation

    The problem
    Processing takes too long or uses more cloud resources than it should.
    What I deliver
    Analysis of existing jobs and proposed changes to compute, data and scheduling.
    The benefit
    More efficient operations. I verify the impact of changes by measuring it.
  • Data streaming from smart devices and sensors

    The problem
    Your devices measure temperature, air quality, or other values, but the data remains separate and you lack a consolidated view.
    What I deliver
    Connections between your devices, sensors, and a data platform, continuous processing of measurements, and data preparation for monitoring current conditions, historical trends, and alerts.
    The benefit
    A consolidated view of measured values and the ability to respond sooner to changes or exceeded thresholds.
  • Microsoft Excel automation

    The problem
    People repeatedly copy, edit, and combine the same spreadsheets. As data volumes grow, files become slow and increasingly difficult to maintain.
    What I deliver
    Automated imports, transformations, and output preparation using Power Query and VBA. If Excel no longer meets your needs, I assess the option of moving data processing to Azure and Databricks.
    The benefit
    Less repetitive work and a lower risk of manual errors. As requirements grow, a path to extending your solution beyond Excel.
  • Business process and document automation

    The problem
    Staff retype data from invoices and other documents and move it between applications by hand.
    What I deliver
    Automated intake of documents, extraction of the data you need and hand-over to spreadsheets or business systems. I design the solution around your process and the applications you use.
    The benefit
    Less admin and faster processing. Uncertain data is checked by a person before it moves on.

Data extraction and integration

Is your data split between an accounting system, orders in an online shop and further information in Excel? Getting a combined overview often means exporting, copying and merging files again and again.

I connect your data sources and set up automatic loading into one shared environment. The refresh frequency depends on how quickly you need the data - for example once a day, or at regular intervals during the day.

How I can help

  • Loading data from databases and business applications.
  • Connecting systems through available APIs.
  • Importing Excel, CSV and other agreed file types.
  • Regular loading of new and changed records, depending on what the source allows.
  • Aligning basic formats and checking that transfers are complete.

I design the exact way of connecting based on what your systems support. If a direct connection is not available, a regular file export can be used instead.

Example use case - illustrative scenario

A company keeps orders in its online shop, invoices in its accounting system and sales plans in Excel. Every week, an employee exports the data by hand and combines it into a single spreadsheet.

I set up regular loading of this data into shared storage. Updated figures are then ready every morning and can be used, for example, to compare orders, invoicing and plan in Power BI.

What you gain

Less time spent moving data around, a lower risk of copy-and-paste errors, and a reliable foundation for further processing and reporting.

Data processing and automated pipelines

Have you collected data from different systems but still need to prepare it manually before each use? Inconsistent formats, duplicate records, and missing values can complicate reporting and everyday work.

I design and build automated processing that cleans, combines, and transforms your data according to agreed rules. I use Apache Spark and Databricks, tailoring the solution to your data volumes and business needs.

How I can help

  • Cleaning data, removing duplicates, and standardising formats.
  • Combining data from multiple sources.
  • Defining rules for missing or invalid values.
  • Calculating metrics, summaries, and other derived data.
  • Scheduling processing steps and managing their dependencies.
  • Handling failures and enabling reruns without creating duplicate outputs.

I agree the processing rules with you. Unclear or invalid data will not be silently deleted or replaced without an explanation.

Example use case - illustrative scenario

A company already collects sales data from several branches in a shared repository. However, the branches use different product names, and some transactions appear more than once. An employee must manually correct the data before preparing reports.

I build a pipeline that standardises product names using an agreed reference list, removes duplicates, and calculates daily sales by branch and product. Records that cannot be processed unambiguously are set aside for review.

The result is a prepared data table for reporting. If a processing step fails, it can be safely rerun once the underlying issue has been resolved.

What you gain

Consistent outputs prepared using the same rules on every run. Less repetitive data preparation and more time to use the results in decision-making.

Data storage and lakehouse design

As your business grows, so does your data. Some remains in business applications, while other information is stored in shared folders and Excel files. Over time, it becomes harder to find current data, understand what individual tables contain, and see how they relate to each other.

I design the structure of your data repository in Azure and Databricks. I define how data will be stored, how source data will be separated from processed data, and which outputs will be prepared for reporting and other uses.

A lakehouse combines storage for different types of data with structured processing for analytics. I tailor the design to your business needs, data volumes, and intended use.

How I can help

  • Designing the storage structure and organising data into logical business areas.
  • Separating source data, cleaned data, and data prepared for analytics.
  • Designing tables, relationships, and historical data retention.
  • Defining naming conventions and data organisation rules.
  • Designing access permissions for different users.
  • Preparing documentation that explains the structure and meaning of the stored data.

I also tailor the scope to smaller businesses. We can start with a specific area, such as sales, and gradually expand the repository.

Example use case - illustrative scenario

A company keeps monthly sales exports, inventory records, and customer data in separate folders. File names are inconsistent, and it is unclear which version should be used for analysis. Older data is sometimes overwritten by newer files.

I design a shared repository with three connected layers. The first retains source data, the second contains cleaned and standardised records, and the third provides tables prepared for analytics.

The design also covers the required history and the relationships between sales, products, and customers. The company gains a clear structure for current reporting and future analysis.

What you gain

A well-organised, documented place for your data. Reports and processing workflows can use a shared foundation that can gradually expand to include more sources and business areas.

Data preparation and models for reporting

Does your sales department report different revenue figures from finance? Does each report show slightly different numbers? The cause may be different sources, filters, or calculation rules.

I prepare a shared data foundation for your reports. I connect the required data, design its structure, and work with you to define what each metric means and how it should be calculated.

How I can help

  • Designing analytical tables and their relationships.
  • Standardising definitions of metrics such as revenue, margin, and customer counts.
  • Defining rules for cancellations, credit notes, dates, and other exceptions.
  • Preparing shared reference data for products, customers, and organisational units.
  • Creating datasets for reports and further analysis.
  • Documenting calculations and validating results against specific examples.

I agree metric definitions with the people who use them. Where departments need different perspectives, I clearly name and document those differences.

Example use case - illustrative scenario

The sales team measures revenue using received orders, while finance uses issued invoices. As a result, each department presents a different total during the monthly review.

I prepare a model that distinguishes order value from invoiced revenue. Together with the company, I define how credit notes, currencies, and reporting periods are handled. Both metrics are then available side by side, with clearly defined meanings.

What you gain

A shared foundation for reporting and clear definitions behind the numbers. Less time explaining discrepancies and an easier way to build new reports.

Power BI reports and dashboards

Do you need to understand how your business is performing but have to search through several files for the answer? Preparing a regular overview can take hours, and the results may already be outdated by the time they are shared.

I create Power BI reports that present important metrics clearly. I start with the questions you need answered and use them to design the content, visualisations, and report interactions.

How I can help

  • Designing reports for business owners, management, and individual departments.
  • Connecting data sources and preparing the required data model.
  • Calculating metrics and comparing results against targets or previous periods.
  • Adding interactive filtering and navigation from summaries to detailed views.
  • Configuring data refreshes according to source capabilities and licensing.
  • Setting up sharing and access, handing over the report, and explaining how to use it.

We can start with a single report for a specific business area. The design will also consider who needs access and how often the data must be updated.

Example use case - illustrative scenario

A business owner receives separate spreadsheets for sales, costs, and targets every month. Comparing performance requires manually reviewing and combining the data.

I create a report showing revenue, costs, and performance against targets. Users can move from the overall picture to individual months, products, or branches to investigate differences. The report also shows when the data was last updated.

What you gain

Important results in one place, with the ability to explore them in more detail. Regular performance reviews become easier, and users can access the information according to their assigned permissions.

Performance and cloud cost optimisation

Are your data jobs taking longer and cloud costs increasing, even though the useful outputs remain much the same? Possible causes include repeatedly processing unchanged data, inefficient file organisation, or computing resources that do not match the workload.

I review your existing processing in Azure and Databricks and propose specific improvements. I focus on changes that can deliver measurable benefits while preserving the correctness of the results.

How I can help

  • Identifying the slowest and most expensive processing steps.
  • Analysing calculations, table joins, and repeated data reads.
  • Proposing a more suitable organisation of data and files.
  • Reducing unnecessary reprocessing of the entire data history.
  • Adjusting computing resources and job schedules.
  • Comparing performance, costs, and output correctness before and after changes.

I prioritise recommendations by expected benefit and implementation effort. Any specific savings are confirmed through measurement.

Example use case - illustrative scenario

A company reprocesses its entire sales history every night, even though only a small proportion of records are new or updated. Processing takes longer as the historical dataset grows.

I assess whether the job can process only new and changed records. I also design a way to capture corrections to older data and compare the outputs with the original solution. I then measure the effect on processing time and resource consumption using comparable runs.

What you gain

A clear understanding of inefficient processing and validated improvements. You will know which changes delivered results and where further work would be worthwhile.

Data streaming from smart devices and sensors

Do you measure warehouse temperatures, office air quality, or humidity in production facilities? Taking measurements is only the beginning. To make use of the data, you need to collect it regularly, retain it, and present it in context.

I design data connections for your smart devices and sensors. I set up the ingestion and processing of measurements so you can monitor current conditions, compare values over time, and receive alerts based on agreed rules.

How I can help

  • Collecting device data through available interfaces or manufacturers' platforms.
  • Combining measurements from multiple devices and locations.
  • Processing measured values, timestamps, and sensor identifiers.
  • Storing historical data to track trends and compare periods.
  • Preparing data for dashboards and visualisations.
  • Configuring alerts when thresholds are exceeded or measurements stop arriving.

I tailor the integration method and update frequency to your devices' capabilities, connectivity, and required response time. The service focuses on the data side of the solution; sensor supply and installation can be discussed separately if needed.

Example use case - illustrative scenario

A company uses temperature, humidity, and CO₂ sensors across several offices. Measurements are viewed separately for each device, with no consolidated view of how conditions change throughout the working day.

I bring the available measurements into a shared environment and prepare data for a dashboard showing current values and historical trends by room. If CO₂ levels exceed an agreed threshold for a specified period, or a sensor stops sending measurements, an alert is sent to the person responsible.

What you gain

A consolidated view of device data, accessible measurement history, and an easier way to identify situations requiring attention.

Microsoft Excel automation

Do you work with similar files every week and repeat the same preparation steps? Copying data, correcting formats, and preparing outputs take time and create opportunities for errors.

Using Power Query and VBA, I automate repetitive workflows so you can complete them by refreshing the data or running a prepared function. I tailor the solution to your files and working practices.

If Excel slows down with larger datasets, processing becomes unreliable, or files become difficult to manage, I also assess moving the data processing to Azure and Databricks. The proposal considers the expected benefits, costs, and ongoing maintenance. Excel can remain a tool for working with prepared outputs, while demanding processing runs on the data platform.

How I can help

  • Importing data from files, folders, and available data sources.
  • Combining multiple tables into a single overview.
  • Cleaning data and standardising names and formats.
  • Preparing recurring summaries and validation outputs.
  • Automating editing, formatting, and exports with VBA.
  • Assessing whether to improve the existing Excel solution or move processing to a data platform.
  • Planning a gradual transition to Azure and Databricks while preserving the required outputs.
  • Providing instructions and clear messages when inputs are missing or invalid.

I consider your Excel version and your company's rules for running macros. During the design, I explain which steps will be automated and where a user will still need to review results or start the process.

Example use case - illustrative scenario

A company receives sales spreadsheets from several branches every month. An employee opens each file, copies the data into a consolidated workbook, and prepares a summary.

I configure Power Query to load files from a designated folder, standardise their columns, and combine the data. Once new files are added, the user simply refreshes the data. If an output must follow a specific format, I add a VBA function to prepare it.

If data volumes and transformation requirements outgrow this solution, I propose moving ingestion and processing to Databricks. Only the required prepared data or summaries are then delivered to Excel. I validate the outputs by comparing them with the original workflow.

What you gain

A simpler, repeatable process for recurring tasks. Less copying and more time to review and interpret the results. If Excel is no longer sufficient, you also gain a plan for further development without having to change your entire way of working at once.

Business process and document automation

Do you manually copy invoice data into spreadsheets or move the same information between applications? Even straightforward administrative tasks can take up a significant part of the working day when document volumes increase.

I design automation covering document receipt, extraction of the required information, and delivery to the destination system. I tailor the solution to your process, applications, and the steps you want to keep under human approval.

How I can help

  • Automatically collecting documents from email or a designated folder.
  • Extracting information from invoices and other agreed document types.
  • Checking required fields, formats, and potential duplicates.
  • Transferring data to Excel, a database, or a business system, depending on integration capabilities.
  • Sorting and archiving documents.
  • Adding notifications, approval steps, and human review of uncertain results.

I first assess the approach using a sample of your documents. Based on the results, I determine which cases can be processed automatically and which require human intervention.

Example use case - illustrative scenario

Supplier invoices arrive in a company inbox as PDF attachments. An employee opens each one, enters the supplier, invoice number, amount, and due date into a register, and saves the document in a folder.

I design a workflow that collects the attachments, extracts the required information, and prepares it for entry. Missing, uncertain, or potentially duplicate data is sent for review. Once the agreed checks and any required approvals are complete, the data is entered into the register and the document is archived.

What you gain

Less manual data entry, faster document processing, and visibility into cases awaiting review. Automation handles repetitive steps while people retain control over decisions that need their attention.

Technologies

The tools I work with

My work is built on Microsoft Azure and Apache Spark. Each solution is tailored to the tools your business already uses.

  • The cloud environment where data solutions run - storage, compute and security.

  • A platform for processing larger volumes of data and building a data lakehouse.

  • Extracting data from invoices, forms and other documents using artificial intelligence.

  • Connecting smart devices and sensors to the cloud to collect measurements and monitor connectivity.

  • Power Query and VBA to automate repetitive spreadsheet work.

  • A connected data platform covering data ingestion and processing through to reporting in Power BI.

  • Automation of repetitive tasks and connections between business applications in a shared workflow.

  • Reports and dashboards for management and individual departments.

Microsoft Azure

The cloud environment where data solutions run - storage, compute and security.

Microsoft Azure is a collection of cloud services that lets businesses store data, run applications and process information without building their own server infrastructure. It can provide a shared foundation for data storage, automated processing, reporting and collecting data from devices.

Businesses can choose services and capacity to match their current needs and gradually expand the solution. Azure also provides tools for managing access, protecting data and monitoring operations. The benefit is the ability to start with a smaller solution and develop it over time. Costs and security also depend on the appropriate design and configuration of the selected services.

Azure Databricks

A platform for processing larger volumes of data and building a data lakehouse.

Azure Databricks is a platform for cleaning, combining and processing data from different sources. It supports recurring workflows that prepare data for reports, analysis or other applications. It is particularly useful when data volumes and processing requirements outgrow manual workflows and individual spreadsheets.

It also helps build a lakehouse: a shared data environment that retains both source and processed data and prepares it for analytical use. Benefits include repeatable processing and the ability to adjust computing capacity as needed. Businesses can gradually develop reporting and other data solutions using the same data foundation.

Azure AI Document Intelligence

Extracting data from invoices, forms and other documents using artificial intelligence.

Azure AI Document Intelligence helps turn document content into data that can be used in other processes. From a PDF or scanned invoice, it can extract information such as the supplier, invoice number, dates, amounts and individual line items. Depending on the documents, processing can use a prebuilt model or be adapted to a specific format.

The benefits are less manual transcription and faster preparation of data for spreadsheets or business systems. The tool handles one part of a wider process, which also needs document intake, validation and delivery of results. Accuracy depends partly on document quality and layout, so uncertain results are reviewed by a person before further use.

Azure IoT Hub

Connecting smart devices and sensors to the cloud to collect measurements and monitor connectivity.

Azure IoT Hub enables communication between devices and a cloud solution. It can receive temperature, humidity, air quality and other measurements from compatible devices. It also supports device identity management and connection monitoring, and can send messages back to devices that support this capability.

The benefit is a shared, managed point for collecting data from multiple devices or locations. Incoming measurements can feed into historical storage, processing, dashboards and alerts. IoT Hub handles device communication; analysis and visualisation are added through other services. The integration approach depends on the capabilities and connectivity of the individual sensors.

Microsoft Excel

Power Query and VBA to automate repetitive spreadsheet work.

Excel can do more than manual data entry and calculations. Power Query can import data from files or other sources, clean it and combine it using a saved sequence of steps. VBA adds automation for tasks such as editing workbooks, formatting and preparing output files.

The benefit is less repetitive copying while keeping the familiar environment employees already use. A prepared workflow can be rerun or refreshed without repeating every transformation from the beginning. If data volumes or processing requirements outgrow Excel, the data processing can move to a more suitable platform while Excel continues to provide the required outputs.

Microsoft Fabric

A connected data platform covering data ingestion and processing through to reporting in Power BI.

Microsoft Fabric brings several parts of working with data into one environment. It lets businesses collect data from their systems, store it, transform it and then use it for analysis and Power BI reports. The platform's components work with a shared data repository called OneLake.

For a business, this can mean fewer separate tools and easier collaboration between the people who prepare and use data. Fabric is worth considering when building a connected solution from source data to management reporting. Its scope and capacity should match the company's needs so that the benefits justify the costs and fit how the platform will be used.

Power Automate

Automation of repetitive tasks and connections between business applications in a shared workflow.

Power Automate connects applications and carries out agreed steps between them. A process can start when an email arrives, a new file appears or a scheduled time is reached. It can then save an attachment, transfer data, request approval or notify the person responsible.

The benefits are less manual movement of information and a consistent approach to recurring tasks. Typical uses include document workflows, request approvals and regular notifications. Integration options depend on the applications, available interfaces and licensing. Extracting data from invoices requires an additional suitable tool, which can form part of the wider automation.

Power BI

Reports and dashboards for management and individual departments.

Power BI turns business data into interactive reports and dashboards. Instead of searching through several spreadsheets, you can track sales, costs, margins or performance against targets in one place. Results can be filtered by period, branch or product and explored from an overall summary down to the details.

Reports can be shared according to access permissions, with data refreshed at agreed intervals depending on source capabilities and licensing. The benefits are a clearer understanding of performance and less manual preparation of recurring reports. Reliable reporting also depends on well-prepared data and clearly defined metrics.

How I work

How we work together

  1. Initial consultation

    In a free online consultation, I find out what you need and how you work with data today.

  2. Solution proposal

    I propose a solution, its scope and how it will be delivered. You receive a tailored quote.

  3. Implementation

    I build the solution step by step and review it with you as we go.

  4. Handover and next steps

    I hand over the finished solution with documentation. If agreed, I continue with further development or support.

Testimonials

What clients say

Translated from Slovak.

  • Working with QSpark helped us replace numerous Excel files with automated Power BI reports. We gained more detailed insights into our performance, faster report preparation and less manual work. Processes that remained in Excel were simplified through VBA automation. We appreciate the practical approach and the choice of suitable solutions for each task. The result is simpler day-to-day work and more time to focus on analysing the data.

    Monika Široká

    Head of Quality Management, Slovnaft a.s.

  • We appreciate QSpark's professional approach. The delivered solution took into account both our needs and the capabilities of our existing systems. Connecting databases and business applications via APIs, together with incremental loading, significantly simplified data collection. Instead of repeatedly loading all the data, we process only new records, which has sped up the process and reduced operating costs.

    Martin Kaša

    Managing Director, Wisecode s.r.o.

  • Thanks to QSpark, we have reduced the manual retyping of data from invoices and other documents into Excel. Processing is faster, checks help catch errors and duplicates, and we have a better overview of our data.

    Viktor Lehocký

    Managing Director, WillCor, s. r. o.

  • QSpark set up the collection and processing of sensor data from multiple devices and locations for us. We got clear data, including measurement history, on which we could build Power BI reports and visualisations.

    Nikolas Németh

    Founder, POWER-BI.CONSULTING, s. r. o.

About me

Bohumil Uhrin

Founder of QSpark, data engineer

I'm Bohumil Uhrin, and I have 12 years of experience in data processing, reporting and automation. I design and build end-to-end data solutions - from extracting data from source systems, through processing, to data warehouses and reports.

I work mainly with Microsoft Azure, Databricks, Fabric and Power BI. But if you insist on Excel, I'll happily open that too. We all have our favourites :-)

QSpark is my own company - from the first consultation to handover, you deal directly with me.

LinkedIn profile

Bohumil Uhrin, founder of QSpark

12 years

of turning data into clear reports, reliable data solutions and less manual work. I use Microsoft Azure, Databricks, Fabric and Power BI.

Contact

Let's talk about your data

Tell me what you need help with. I'll get back to you and we can arrange a free initial online consultation.

Send an email

Email
bohus.uhrin@gmail.com
Phone
+421 915 241 878
LinkedIn
Bohumil Uhrin