Data Science for Public Policy (MS) postgraduate personal statement example

Personal statement example

Each August I rebuild a spreadsheet that decides which children are collected at 7.40 and which at 8.05. I work as a rota and data coordinator for a small contractor that runs home-to-school transport for a county council, and the spreadsheet holds pupil postcodes, eligibility codes, vehicle capacities and driver hours. When the council changed its walking-distance threshold two years ago, I watched a policy sentence become forty new rows, three extra minibuses and several anxious phone calls from parents. That experience is why I want to study data science for public policy: I have seen how rules travel into data, and I want the skills to examine that journey properly rather than simply process the result.

My degree in Geography gave me a foundation in quantitative methods, including regression, spatial statistics and survey design. For my dissertation I used publicly available GTFS timetable data to estimate how many residents of two neighbouring towns could reach a GP surgery by bus within thirty minutes on a weekday morning compared with a Sunday. I taught myself to work with the timetable files in R, joined them to census output areas, and calculated travel times with a routing package. The hardest part was not the code but the definitions: whether to count waiting time, how far people realistically walk to a stop, and what to do with services that ran only on school days. I documented each choice and ran the analysis under two walking assumptions, which showed that the gap between weekday and Sunday access was large under either. My supervisor's main criticism was that I had not tested whether timetabled services actually ran on time, which is a fair limitation and one I would now address with real-time data where it exists.

At work I have built on those skills in modest ways. I replaced a manual process for checking driver hours with a set of spreadsheet formulas and, later, a short Python script that flags shifts breaching our limits before rotas are sent out. My manager now uses it weekly. I also produce a monthly summary for the council contract officer showing late collections by route. Preparing it taught me that a clean chart can hide unevenness: one route looked acceptable on average while a single morning run was late most days because of roadworks. Since then I always look at the distribution before the mean.

Outside work, I sing in a community choir and help organise its concert ticketing, and I help my grandmother with online forms for her pension and council tax. The second is not research, but it has made me attentive to how services designed around data can be confusing for the people using them. A form that asks for a reference number she has never seen is a data problem as much as a communication one.

My reading has followed these interests. Cathy O'Neil's Weapons of Math Destruction made me think about how automated scoring can entrench disadvantage when its workings are hidden, and Virginia Eubanks's Automating Inequality gave concrete examples from public services in the United States. I have also worked through introductory material on causal inference, particularly difference-in-differences, because I would like to know whether the walking-distance change actually altered attendance, not just bus numbers.

I am applying for postgraduate study because I have reached the limit of what I can teach myself reliably. I want rigorous training in statistical modelling, machine learning and causal methods, alongside the policy context needed to ask good questions about them. I bring practical experience of messy administrative data, a habit of writing down assumptions, and an understanding of what a policy change looks like from the operational end. I hope to work afterwards as an analyst in local government or transport planning, helping decisions rest on evidence that is clearly explained.