University of Oxford

Imperial College London

University of Cambridge

CEO & Co-Founder @ The Data Analysis Bureau  -  BIMA AI Council Member

London, London, United Kingdom

The Data Analysis Bureau is a data science and data engineering innovation company specialising in developing and building machine learning solutions.Harnessing the latest advances in advanced analytics, machine learning, and cloud computing, we deliver cutting edge data driven technology solutions providing actionable insights to help our clients take advantage of new opportunities. Our value-oriented approach means we focus on quality to deliver results back to the business rapidly.We also work with trusted partners to embed machine learning and advanced analytics into both industry agnostic and vertical specific products. We have to date secured just over £1.3m of funding for such initiatives.Through our T-DAB lab, the Innovation Sandbox, we are active in collaborating with academia in machine learning research and development, contributing actively to advancements in natural language processing, reinforcement learning and intelligent control, time to event prediction, and anomaly detection.We believe in AI for good and that smarter data means better decisions.

The Data Analysis Bureau

Chief Executive Officer

I am the interim Chief Scientist at Get-Optimal, advising on the build of a deep learning driven SaaS platform that optimises job adverts for diversity and inclusion.

Get-Optimal.com

Chief Scientist

As Chief Scientist my role is to advise on the design, development, and delivery of core technologies that underpin the functioning of AI.Market as the leading marketplace for Data and AI. My remit covers leading  development of the platform to provide the data storage and processing, model development, and data and AI ethics tools that will allow actors in the market to confidentially share data, ethically develop models, and transact these assets securely.

ai.market

BIMA (British Interactive Media Association)

Member AI Council

Part of the the new data science arm for the forecasting team at London ambulance service. I am working on improving legacy time series and forecasting models and moving them in to more powerful environments (from excel/VBA to R Server). I will be looking to applying more powerful methodologies including machine learning to new areas of the business.My current major project is to build predictive models for bariatric ambulance demand and resource allocation. This suite of models will predict London specific obesity related phenotype demographics,  how these ambulance demand, how demand is driven spatially, and finally how to allocate resources accordingly.Skills used:- multivariate regression- mixed effects modelling- time series- forecasting and predictive modelling- resource allocation modelling- R, SQL, Excel

London Ambulance Service

Data Scientist

I implemented the early stages of a project for the Vascular Physiology Unit at UCL to advise on the analysis required for an observational study of 450 young adults in London from the long running Avon Longitudinal Study of Parents and Children (ALSPAC). This consisted of anthropometric and cardiovascular data linked to previous ALSPAC data sets to create a longitudinal data set from birth. This entailed:- Leveraging clinical knowledge of the VPU to help construct testable hypotheses- Advising on appropriate statistical methods and modelling - Advising on limitations of the data and of techniques - key to helping the client draw robust results- Data cleaning - Structuring of the dataset- Initial investigation and inspection of data- Statistical modelling - linear regression, logistic regression, generalised linear modelling, (generalised) linear mixed effects modelling- Analysis carried out in R and R server- Visualisation in GG plot 2- Interpretation of statistical results to the stakeholder (VPU research group) in a relatable manner with context appliedI also consulted on statistical work for research by the VPU into the effects of common anti-inflammatory drugs on the protective function of high density lipoproteins (HDLs) in human vascular systems. This involved:- Rapidly assessing methods used for work previously submitted and rejected for publication- Assessing the criticisms of the reviewer- Inspecting the data - identifying erroneous statistical assumptions made in previous analysis- Providing to the stakeholder (VPU research group) an alternative analytical approach using more advanced and robust statistical techniques correcting for previous errors- Clarifying hypotheses- Re-analysing the data in R using  general linear and generalised linear mixed effects modelling- Visualisation in GG Plot 2- Interpretation of statistical results to the stakeholder (VPU research group) in a relatable manner with context applied

Vascular Physiology Unit (UCL)

Consultant Statistician

I performed clinical audit work of the performance of cardiologists and cardiology departments carrying out percutaneous coronary interventions (PCI) and trans-catheter aortic valve implementation (TAVI). This involved working heavily with numerous stakeholders; cardiology consultants, professional cardiology bodies (BCIS), as well as other public stakeholders such as HQIP.The statistical work primarily involved mixed effects modelling and risk modelling, implemented in R server using big data collected through the NICOR web portal. These were then used to produce quarterly and yearly reports which are disseminated to both clinicians, hospitals and also made available to the public. I was also involved in reconstructing and re-developing some of the computing and statistical tools implemented at NICOR after a period of restructuring. This involved work in R, R server, R Shiny, and cooperation with data developers (primarily regarding the format and structure of extracts from the ETL pipeline).Key skills and responsibilities:- Ran two clinical audits- Very large dataset (>500,000 operations and 168 fields) of population level data for England, Wales and Northern Ireland- Reconstructed and re-developed damaged/lost legacy R scripts- Implementation of automated data cleaning routines in R Server- Cleaning of data extracts provided by data developers from the ETL pipeline- Coding of statistical analyses in R server - GLM, linear mixed effects, elementary statistics (means, medians, percentages, counts etc.)- Outlier analysis - funnel plots- Working in conjunction with stakeholders (Clinical leads, technical steering groups) to design and code new logics for dataset subsetting and report statistics- Visualisation in GG plot 2- Automated generation of reports using R server, Knitr and LaTex- Coding automated report elements based on conditional rules- Editing and version control of data and code dictionaries in Lotus Notes

Contractor - Data Scientist (Statistical Programming)

This project primarily involved implementing complex modern statistical modelling techniques to build a new work load formula for financial allocations to CCG areas for primary medical care. The work and formulae produced were considered to be improvements on the previous Carr-Hill formula and will be used in current and future financial allocations as well as GP contract negotiations. The analyses were implemented in R and concerned complex mixed effects modelling as well as modern imputation techniques, notably multiple imputation by chained equations (MICE). Work was independently quality assessed by consultants from Deloitte.Key skills and responsibilities:- Advisory role on how to approach the project from the perspective of statistical modelling- Advisory role on potential pit falls of the dataset from the perspective of statistical modelling- Providing advice on appropriate types of statistical methods and modelling - Advising on potential limitations of the data and of techniques - key to helping the client obtain robust results- In conjunction with the wider analytics team and independent expert, assessing the appropriateness of possible predictor and response variables and of any likely transformations/corrections- Data cleaning - time/date format/outliers/illegal values/assessing for incidence of missing data- Initial exploratory analyses - distributions, basic predictor-response variable relationships- Statistical modelling of GP workloads using linear regression, GLM, and linear mixed effects models- Model criticism and simplification- Investigation of imputation of missing data on the dataset and models (by list-wise deletion, zeros, means, and multiple imputation by chained equations [MICE])- Implementation of MICE- Working with econometricians (from Deloitte) present to QA the work- Visualisation in GG plot 2- Interpretation of statistical modelling results for technical and non-technical elements of steering groups and stakeholders

NHS England

Consultant - Data scientist (Statistical Programming)

I provide contracting and consulting services in data analysis, statistics and coding in R. With years of experience of working with large and complex data sets in an empirical research background, I have the experience to help my clients prepare, analyse, and interpret their data such that they can have the utmost confidence in - and understanding of - the results.

Eric Topham Limited

Director

Lead teaching assistant on graduate and undergraduate courses in statistics and R.Teaching assistant on undergraduate behavioural ecology and field ecology courses.

Undergraduate/Postgraduate Teaching Assistant

Involved re-connecting and networking with alumni of Merton College, Oxford on behalf of the college development office in order to raise money for student hardship funds and university fee waivers. Achieved the highest rate of donation and highest average donation per call, while further developing my interpersonal skills.

Merton College, Oxford

Alumni Donations and Liaison

My doctoral research concerned new methods of quantifying functional ecological diversity, the mechanisms that maintain biodiversity predators (particularly when subjected to environmental variability), and how changes in diversity of prey species impacts the contributions of predator species to ecosystem functioning. My experiments used environmentally collected microbial and protistan systems to construct communities and bridge the divide between field and laboratory studies. I am currently working on publishing this work.Key skills and techniques:- Advanced statistical linear and non-linear modelling techniques (linear regression, multiple regression, GLM) in R- Application of novel and bespoke statistical models to determine contribution of species to ecosystem functioning- Parametric and non-parametric statistics- Visualisation in GGPlot2- Development of novel mathematical clustering techniques to classify functional groups of organisms- Experimental design- Scientific writing- Public speaking and presentation skills- Laboratory techniques (bio/enzymatic assays, sequencing, PCR, cell culture, sterile technique, microscopy)- Management and teaching

Dphil Student (PhD) in Zoology

Providing teaching support during undergraduate teaching laboratory sessions in Microbiology and Evolution. This included being entirely responsible for running sessions with approximately sixty students.Providing one-to-one tuition in statistics for third year undergraduates.

Undergraduate Teaching Assistant

Managed a small team of research assistants in collecting and analyzing species distribution and abundance data to subsequently carry out assessments for the IUCN Red List for butterflies.

Research Assistant

Provided statistical analysis of departmental data. Methods used included contingency table analysis and linear modeling. Analyses implemented in R.

Reconnecting with Alumni of Fitzwilliam College, Cambridge to raise money for student hardship funds.

Fitzwilliam College, Cambridge

Student Fundraiser

Dive preparation, safety in and out of water safety, and dive leadership. Also worked on providing in water educational tours for school groups teaching basic marine biology.

Diving Center La Sirena

Divemaster

Responsible for aspects of dive preparation, safety (in and out of the water), and assisting with instruction.