Data engineering postgraduate personal statement example

Personal statement example

My final-year project began with a simple annoyance: I could see live bus arrival predictions on my phone, but I could not easily find out whether the buses on my route were usually late. Open data existed, but it arrived as awkward XML feeds that changed shape without warning. I wrote a Python service that polled the feed every two minutes, stored raw responses before touching them, and only then parsed records into a PostgreSQL schema. Keeping the raw layer separate turned out to be the most useful decision I made, because when I discovered that I had been misreading a timestamp field for three weeks, I could reprocess the history rather than start again. That experience taught me more about data engineering than any single module: pipelines are judged on how well they survive change, not on how elegantly they run once.

My degree gave me the foundations I want to build on. I enjoyed databases and distributed systems most, particularly working through normalisation and query planning until I could predict why an index helped. Algorithms and complexity were harder for me, and I spent a lot of time on practice problems to reach a solid mark; I would rather say that plainly than pretend everything came easily. Outside the syllabus I have worked through material on SQL window functions and on Spark, and I have been reading Designing Data-Intensive Applications, which has reframed how I think about replication and about the trade-offs between batch and stream processing. The chapters on encoding and schema evolution made sense to me precisely because I had already been bitten by a changing feed.

Since graduating I have worked as a junior IT support and reporting assistant for a housing association. Much of the role is routine: resetting accounts, fixing printers, answering questions about a case management system. The reporting side is where I have learned the most. Our monthly repairs figures were assembled by hand from three spreadsheet exports, and small differences in how people typed contractor names produced totals that never quite matched. Working with a colleague from the repairs team, I rebuilt the process as a set of SQL views over an imported copy of the data, with a lookup table for contractor names that she maintains because she knows the business context and I do not. The report now takes minutes rather than most of a morning, and errors are easier to trace. It is a modest improvement to one process, but it showed me how much of this work is agreeing definitions with the people who use the numbers.

I want formal training because I have reached the limits of what I can pick up in the gaps of a support job. I can write reliable Python and reasonable SQL, but I have only read about orchestration tools, warehouse modelling and cluster computing rather than used them under pressure with data too large for one machine. I am also aware that I have never had to think seriously about governance, lineage or handling personal data at scale, which matters given that the records I touch now concern people's homes. A postgraduate course would let me work on larger, messier datasets with proper feedback on my design choices, and I would like a project that involves streaming data, ideally something transport related so I can return to the questions my undergraduate work left open.

In the longer term I want to build and maintain the pipelines that analysts and machine learning teams depend on, preferably in the public or third sector. Outside work I run the fixtures and availability spreadsheet for my cricket club, which is a reminder that most data problems start with people entering things inconsistently.