Transforming Glasgow | Data Science & Analytics Lead | 2016-2018

About Link to heading

At the end of the 2 year Future Cities programme - which functioned independently from the city government - I was brought in-house to work on the council’s $250m transformation programme. I was remanded to the team tasked with finding data-led solutionst to financial and organisational challenges.

My previous role had set me up well - with a good network of the data stakeholders and knowledge of how to work with and around the beaurocratic challenges that come along with a large political institution.

Over the following 2 years I worked on more than a dozen fascinating projects across the entire domain of the council’s work - and my work directly impacted countles services and led to a considerable quantifiable impact in the city which I’m extremely proud of.

Privacy & Compliance

One of the most compelling elements to the work was working with extremely sensitive datasets - guarded by some real and some more perceived privacy concerns in the nascent era of GDPR.

I developed a close working relationship with the privacy and compliance leadership - learning in detail the regulatory context and allowing me to make rational and positive cases for internal data sharing and unlocking valuable data resources inside the organisation.

Ultimately the impact I was able to have was a result of this cross disciplinary approach and an excellent model for innovation in a tightly regulated environment.

Benefits auto entitlement

This project was significant as it started small, but ended up having a great impact.

There was a programme run by the city to hand out food benefits to certain demographics. The process required these beneficiaries to apply and included means testing - which is costly and time consuming to administer.

I was able to bring together multiple data sources - including the central government dataset for means testing benefits recipients - and use probablistic data matching to automatically determine and pay out the benefits to the elligible beneficiaries.

This project raised percent of those who were elliglbe vs those who we covered from ~75% to 96%, meanwhile reducing the cost of administration by nearly 95%.

As an unexpected bonus - there were other government programmes that were tied to these food benefits - so we ended up inadvertantly bringing just under $10 million new funds into the city with this programme, in addition to the cost reduction and additional grants.

Freedom of Information Requests

This was another quick win with a big impact. The council spent significant amounts of budget each year servicing freedom of information requests from the public and civil society.

The system for submitting these requests was archaic - and a lack of knowledge resources on both the public and public service side meant a lot of time was wasted on requests that would ultimately be denied.

I was tasked to look at this and see what we could do with respect to publishing the most frequently requested datasets - processed and ready for consumption - so the team traiging the requests could point to these datasets and significantly reduce the cost and time spent on them.

My approach was two fold - firstly to do some data cataloguing around the main valuable data resources inside the organisation - to help build a more intuitive and shared understanding of who had what and what it was.

The second was to do some NLP text procesing on the FOI requests themselves to build a quantatative model of what people were after.

We ended up taking this a bit further with setting up API’s to automatically publish the high-demand datasets. I particularly enjoyed working on this project because it had a big financial impact despite the original idea being so obvious - it could have been easily overlooked.

Impact Link to heading

  • Applied statistical methods to identify gaps in elligibility for single mothers receiving support income, and raised entitled single-mother households from 70% - 95% while reducing cost of administration.
  • Portfolio of data science projects resulted in $10m of realised savings and $40m identified savings.
  • Built up internal network of data owners & analysts to facilitate data-led decision making at the organisational level.

References

Technologies