Work
Forecasting Pipeline & Monitoring App
Automated forecasts for sales data: preparing and validating the data, training and comparing time series and ML models, and surfacing deviations instantly in a monitoring app.
Go to project pageTimmy Lu · Data Analyst & BI Developer

I bring together data analysis, business intelligence and automation: I prepare data so people can trust it, build dashboards teams make decisions with, and automate whatever repeats – more and more often with AI.
Work
Automated forecasts for sales data: preparing and validating the data, training and comparing time series and ML models, and surfacing deviations instantly in a monitoring app.
Go to project pageWork
An LLM workflow checks product data from several online shops for missing or incomplete attributes – an analysis that used to be done by hand now takes a fraction of the time.
Work
KPI cockpits for e-commerce and sales with plan and prior-year comparisons, plus a dashboard for influencer campaigns – from the first requirement to a report that is up to date every day.
Go to project pageWork
Jira data flowing into the data warehouse automatically: queried via REST API and JQL, loaded incrementally, staged and merged into Exasol – as the basis for Qlik reports.
Work
Connecting external data sources via REST APIs, cleaning and standardizing the data – and making it usable in ERP, data warehouse and BI reports.
Work
Learning to code with a real robot: making it walk, talk, play music, blink its LEDs and adding sensors – in internal workshops and at school events like Girls’ Day.
Go to project pageUniversity
Purchase prediction with graph neural networks: customers and products as a temporal graph, link prediction with temporal GNNs and graph attention.
Go to project pageUniversity
Team project
Building management software developed by a team of three. My part: user and task management – from the interface through gRPC services down to the database, including file storage, comments and a custom date picker.
Go to project page


Own projects
Work in progress · simulation only
A trading system that tests strategies on real market data: data pipeline, backtester, paper trading bots, model tournament and web dashboard. Runs as a simulation only.
Go to project page



Design mockup with invented example numbers – not real results.
2021
Analysis tool for the daily log files of a customer registration on Windows devices: counts events per device model, calculates error rates (above 1 % marked red), draws charts – with a user interface and an automatic weekly report by email.
2022
Test program for notebook batteries: switches Wi-Fi sockets on and off automatically so charging cycles run unattended, and records the measurements as a chart and CSV. My part: the interface with charts and dark mode, and the user guide as a web page.
2025
Calculates from inbound lists how many pallets arrive at the warehouse per calendar week – including pallet factor and two-person handling, as a Python script running inside the database.
2024
Market research and evaluation of marketplace middlewares as an alternative to the current provider: defining criteria, comparing providers and their marketplaces, preparing a recommendation.
2025
Why do sales of a product group go up? Search interest compared with sales, traffic sources and regions analysed – and the result prepared for the business team.
since 04/2026
Essen
04/2025 – 03/2026
medion GmbH, Essen
08/2021 – 03/2025
medion, Essen
09/2021 – 03/2025
FOM University of Applied Sciences, Essen

Hip-hop and urban dance – training, performances and competitions.
Piano since childhood, plus guitar and my own songs in FL Studio – somewhere between indie and bedroom pop.
Self-hosted game servers and mods on my own cloud VM.

I’m open to full-time roles in data analytics, BI and AI engineering – remote preferred, hybrid is open for discussion.
