Data your management can rely on.
I build data platforms, data models and dashboards: from connecting the sources to the figure in the management report. Freelance and remote, in German or English.
Sonja Armbruster, Senior Data & Analytics Engineer based in Frankfurt. Working with data since 2004: 18 years at Commerzbank, then projects for AstraZeneca, CLAAS and Groz-Beckert.
- Role
- Senior Data & Analytics Engineer
- In data since
- 2004
- Industries
- Banking, pharma, agricultural machinery, engineering
- Tools
- Tableau, Power BI, Databricks, SQL, Python
- Way of working
- Freelance, remote
- Languages
- German, English
- Based in
- Frankfurt am Main
What I build for you
From the source system to the dashboard: I take on the whole pipeline, or the part your team is missing.
Data platforms and pipelines
Bringing scattered sources together on one platform, with Databricks, PySpark and Spark SQL, Delta Lake and a medallion architecture. I connect sources such as Funnel.io, Matomo or Salesforce Marketing Cloud through their REST APIs, and the runs update incrementally.
Data models and data quality
A clean data model in which every metric is defined once, so every department works with the same number. Plus quality checks, reconciliation, monitoring and documentation.
Dashboards and reporting
Management dashboards and self-service BI in Tableau, my strongest tool, as well as in Power BI and MicroStrategy. Built on datasets prepared for reporting, so reports load fast.
Analysis and predictive models
Segmentation and predictive models such as random forest and XGBoost, for example for predictive maintenance or cross-selling, with results that management and business teams can use directly.
Selected projects
Long-running projects in agricultural machinery, engineering, pharma and banking.
-
CLAAS
Senior Data Analyst & Data Platform Engineer
Marketing data, agricultural machinery2024–2025
Starting point
Nine isolated marketing data sources with no shared keys. The data flowed unfiltered straight into Tableau.
What I built
A marketing data platform in Databricks that combines Funnel.io, Matomo and Salesforce Marketing Cloud with quality checks, plus dashboards in Tableau and Power BI on the customer journey and on the executive team's KPIs.
- 6M+ records a day
- 7 sources, 20+ countries
- Runs in 1–2 minutes
-
Groz-Beckert
Data Engineer & Analyst
Production data, engineering2022–2024
Starting point
10 million sensor records with no structure and no quality control. The goal: find the patterns behind machine failures.
What I built
IoT data pipelines on AWS with systematic validation, monitoring and logging, predictive maintenance models, and reports for management and production.
- 10% less unplanned downtime
-
AstraZeneca
Business Intelligence Analyst
Business intelligence, pharma2022–2023
Starting point
More than 20 reports a day, built by hand, with no automation and no common standard.
What I built
Automated sales and audience analytics with SQL and Python, standardised the reporting workflows, and built management dashboards in MicroStrategy and Power BI.
- 20+ reports automated
- Management dashboards
-
Commerzbank
Quantitative Risk Analyst, then Data Analyst in customer analytics
Risk and customer analytics2004–2022
Starting point
ECB and BaFin requirements for risk models and reporting, and later the migration from the data warehouse to a data lake.
What I built
Scoring and rating models, auditable data pipelines for stress tests and risk reporting, the migration to the Hadoop data lake, customer behaviour dashboards in Qlik Sense, Tableau and Power BI, and a cross-selling analysis.
- 18 years of regulated data
- €120M business potential identified
How I work
18 years of bank data under ECB and BaFin requirements shape how I work: every number has to be traceable.
Principles
- Freelance and remote, in German or English
- I work in your systems, with access you grant and can revoke
- Metrics defined in writing, results reconciled with your existing reports
- Documentation and handover, so your team can keep building on its own
Education and training
- MBA, Quantitative Methods and Business Informatics
- Tableau Desktop and Tableau Visual Analytics
- Neural Networks and Deep Learning
- Machine Learning with Python
Contact
Tell me briefly what it's about. I'll get back to you as soon as I can.
Typical requests
- Build a data platform, or clean one up
- Align metrics across departments
- Build new dashboards in Tableau or Power BI, or replace old ones
- Strengthen an existing data or BI team
Prefer email? info@sonjaarmbruster.com