Edinburgh RSS local group
Meetings 2011

Wednesday 23rd February 2011
Talk about revolutions: the changing face of data dissemination in the 21st century
Dr Steve Rogers (ONS Data Visualisation Centre)
Room G.07a, Informatics Forum, University of Edinburgh, 10 Crichton Street
12.30pm - 4.30pm
The event is free to attend, however places are limited to 50 therefore registration is essential.

Dr Steve Rogers from the ONS Data Visualisation Centre will present an engaging talk charting historical and recent developments in Data Visualisation and other related sciences such as Visual Perception and Information Technology that can be used to improve the way in which statistics are communicated. This will be followed by an informal Q&A session in which Steve will further demonstrate the tools developed by the Data Visualisation Centre and discuss specific issues or areas of interest related to the presentation of statistical information.

The shift in data dissemination from traditional paper to web-based publication is the latest challenge in a history of revolutions that have shaped the way National Statistical Institutes deliver statistical information to the world. While inventions such as time-line charts, line and bar graphs, and the steam-press, opened the door for presenting statistical information to a much wider audience in a 'meaningful' way, the World Wide Web and Web 2.0 marks the beginning of new opportunities to unlock and present even more information latent in data. Guided by other revolutions such as our understanding about how people perceive information in both static and moving images, here, we present some of the ways the Office for National Statistics is meeting the challenge of data dissemination in the 21st Century.

Tuesday 8th March 2011
Asset Risk Modelling in Scottish Water
Robert Murray (Scottish Water)
ICMS, 15 South College Street, Edinburgh EH8 9AA
Meeting starts 6pm, tea and coffee from 5.30pm

Scottish Water is a Government owned corporation providing a vital service to Scottish Society. That is, the abstraction, treatment and consistent supply of clean safe drinking water to every home and business across Scotland as well as the collection, treatment anddisposal of resulting wastewater in a manner that protects the environment. The company's asset management strategy sets out the intended long-term approach to the management of its broad and diverse asset stock; comprising approximately 97,000km of water and sewer pipes, 300 water treatment works and 1,800 waste water treatment works. This strategy includes two key themes. The first being, the development of objective repeatable quantitative methodologies to understand how asset criticality and risk changes over time. The second being, the delivery of capital and operational activity proactively to ensure sustainable and efficient service to customers. Scottish Water has established an in-house team of statisticians to model how, when and where investment should be targeted to maximise benefit to society. The talk will outline in more detail the work that the team does.

Monday 4th April 2011
Measuring the UK's progress: some reflections on holding a national debate on well-being and on what matters to people

Paul Allin (Office for National Statistics)
Melting Pot, Rose Street, Edinburgh
Meeting starts 6.30pm, tea and coffee from 6pm

At the end of November 2010, the National Statistician accepted an invitation from the Prime Minister to develop wider measures of national well-being. The two main strands of this were to hold a national debate on measuring well-being, which runs until 15 April, and to work up questions on subjective well-being to add to ONS's Integrated Household Survey from April. In this talk we will review progress and the emerging findings from both strands, and outline possible future developments. There is also be an opportunity for the audience to say 'what matters' to them. The talk will also consider the ambiguity of 'national' in this context, as on one hand, this work aims to find broader measures that go beyond GDP and other statistics from the UK National Accounts while international organisations (EU, OECD, UN) are also pressing for the development of measures at Member State level. On the other hand, we need to understand and respect the devolved nature of most, if not all, of the policy areas in which well-being will become increasingly important, as well as the roles of governments beyond Westminster in promoting well-being and progress within Northern Ireland, Scotland and Wales. For more information on the ONS programme of work, and to take part in the debate outside of this meeting, please go to www.ons.gov.uk/well-being.

Paul Allin is a government statistician and director of the Measuring National Well-being programme. He has worked for the Office for National Statistics since 2002 and is also responsible for the ONS's social reporting, including Social Trends. Paul started work as an assistant statistician on the national accounts in the then Central Statistics Office in London, moving to work in several government departments and agencies in London and in the north west of England. From 1994 to 2002 he was the chief statistician and head of the Social Policy Unit at the Department for Culture Media and Sport (DCMS). He continues to be a member of the advisory boards of the journals Cultural Trends and Arts Research Digest. Paul is now based in South Wales and chairs the RSS local group there.

Tuesday 17th May 2011
What role should formal risk-benefit decision-making play in the regulation of medicines?
Prof Deborah Ashby (Imperial College London)
ICMS, 15 South College Street, Edinburgh EH8 9AA
Meeting starts 6pm, tea and coffee from 5.30pm

The regulation of medicine requires evidence of the efficacy and safety of medicines, and methods are well-developed to deal with the latter and to a lesser extent the former. However, until recently, assessment of risk- benefit especially in relation to alternatives has been entirely informal. There is now growing interest in the possibilities of more formal approaches to risk-benefit decision-making. In this talk, we review the basis of drug regulation, the statistical basis for decision-making under uncertainty, current initiatives in the area, and discuss possible approaches that could enhance the quality of regulatory decision-making.

Tuesday 11th October 2011
Prof Colin Aitken and Amy Wilson (School of Mathematics, University of Edinburgh)
Evaluation of evidence relating to traces of cocaine on banknotes
ICMS, 15 South College Street, Edinburgh EH8 9AA
Meeting starts 6pm, tea and coffee from 5.30pm
Download poster(PDF file)

Banknotes can be seized from crime scenes as evidence of illicit drug use / dealing. Mass Spec Analytical Ltd. (MSA), an analytical chemistry company in Bristol, have developed a technique to analyse the levels of drugs on banknotes. They have collected data from banknotes seized in criminal investigations, as well as from banknotes from the general circulation. For each note tested, data are available on the peak area for cocaine. Some of the principles involved in analysing these data will be introduced, focussing on the likelihood of the data under each of two propositions: that a set of seized banknotes is associated with drug crime, and that these banknotes are from the general circulation. The aim of the research is to determine and evaluate the associated likelihood ratio. The problems faced in the calculation of the likelihood ratio are discussed. Models based on Hidden Markov Models with autoregression will be described, with some preliminary results.

Tuesday 15th November 2011
The Scottish Health Survey
Catherine Bromley (Deputy Director of the Scottish Centre for Social Research, and director of the 2008-11 Scottish Health Survey)
Dr Linsay Gray (Investigator Scientist, MRC/CSO Social and Public Health Sciences Unit, Glasgow and Lecturer (Honorary), College of Medical, Veterinary and Life Sciences, University of Glasgow)
ICMS, 15 South College Street, Edinburgh EH8 9AA
Download poster (PDF file)
Meeting starts 6pm, tea and coffee from 5.30pm



Scottish Health Survey - past, present and future (Catherine Bromley)
The Scottish Health Survey is an important research resource for Scotland. It provides both cross-sectional and longitudinal analysis opportunities on a range of topics, with a time series dating back to 1995. It covers general health and mental wellbeing, the burden of cardiovascular and respiratory disease, and behavioural risk factors (smoking, drinking, physical activity, diet, obesity). In addition to self-reported measures, it includes objective biomeasures collected via a nurse visit (blood pressure; lung function; waist/hip circumference; saliva, urine and venous blood samples). The survey underwent a number of methodological changes in 2008, and is about to undergo more changes from 2012. This paper will outline the current analysis potential of the study and set out the implications of the changes being introduced for future analyses.

Socioeconomic gradients in coronary heart disease - the relative role of lifestyle (Linsay Gray)
Cardiovascular disease (CVD) is known to be strongly associated with socio-economic status (SES). Aspects of lifestyle such as cigarette smoking, alcohol consumption, physical activity, diet and body mass index impact on health and are also independently linked with SES, and thus may drive the CVD-SES correlations. Knowledge of the relative impact of individual and combined contributions of lifestyle factors is essential for tackling socio-economic inequalities in health, but this has not been well quantified. We address this by applying Cox proportional hazards regression using the relative index of inequality in prospective data created by record linkage of Scottish Health Survey 2003 data to hospital admission and death records.

Tuesday 13th December 2011
Joint meeting on financial statistics between the RSS Edinburgh local group and PRMIA (Professional Risk Manager's International Assocation)
ICMS, 15 South College Street, Edinburgh EH8 9AA
Meeting starts 6pm, tea and coffee from 5.30pm

This meeting will include talks by two speakers:

Dr David Redfern (Senior Analyst in Model Research, Barrie and Hibbert)
Low Discrepancy Numbers in Financial Stochastic Modelling
Uncertainty in financial markets is often modelled using stochastic differential equations. There is no shortage of financial instruments in current use which are too difficult or time-consuming or just plain impossible, to employ in analytic modelling. Consequently, Monte Carlo methods are in high demand. Such methods use samples taken from one or more distributions to produce many realisations of possible financial futures. These can then be analysed to produce expected values, confidence intervals and/or distributions of the particular financial quantities of interest. The most common Monte Carlo schemes use random sampling, but in this presentation I will discuss the use of deterministic low discrepancy sequences for the sampling. Such sequences (including Sobol sequences) have been shown to drastically improve the efficiency of Monte Carlo techniques (reducing run time and/or improving accuracy). Compared to other applications, financial modelling tends to involve higher dimensional problems which can prove quite challenging and I will outline some of the pros and cons of using low discrepancy sequences in financial settings.

Dr Galina Andreeva (University of Edinburgh Business School, Credit Research Centre)
Statistical discrimination in consumer credit and insurance: implications of legal compliance
On March 1, 2011 the European Court of Justice ruled that "taking the gender of the insured individual into account as a risk factor in insurance contracts constitutes discrimination". This presentation looks into aspects of statistical discrimination between good and bad risks which in areas of retail banking and insurance is often based on personal information, such as gender or age. Using numeric examples, the discussion demonstrates possible implications of legal restrictions on the use of certain information in risk assessment. Such implications include deterioration of predictive accuracy of risk models, and reduced access to credit for good risks. Additional difficulties in eliminating the effect of prohibited variables arise from a highly inter-related nature of predictors or risk factors in risk assessment models. Alternative scenarios are considered and difficulties in achieving ethical decisions are discussed.