Tuesday 25 February 2014, 6-7pm
Census 2011: Results so far and innovative ways to access data
Sandy Taylor (Statistician from National Records of Scotland) and Justin Hayes (Census Support Service Manager, Mimas, University of Manchester)
ICMS, 15 South College Street, Edinburgh EH8 9AA
Tea & coffee available from 5pm, and the Annual General Meeting of the Edinburgh local RSS group will be held at 5.30pm; all are welcome to attend
Scotland's Census 2011, Sandy Taylor
The session will: provide some detail of the processes that lie behind the processing of Scotland's Census 2011 (e.g coverage and imputation); highlight some of the key results so far; and, offer an introduction to the Census Data Explorer - National Records of Scotland's key online resource for making the results from Scotland's Census available to users.
New and easier ways of working with aggregate data and geographies from UK censuses, Justin Hayes
The aggregate outputs from UK censuses provide detailed, high quality information on a wide range of the demographic and socio-economic characteristics of people and households across the UK. They are, however, very large and complex datasets, which can make them difficult to understand and use effectively. This presentation will describe and demonstrate how new methods of managing and providing access to the aggregate outputs from the UK 2011 Census via the InFuse interface (infuse.mimas.ac.uk) have made them easier to work with to exploit the valuable information they contain. InFuse offers several unique features including:
- Searches for data across the entire collection of UK aggregate outputs by selecting combinations of variables rather than searching through hundreds of traditional tables.
- Selection of data for equivalent geographical areas across the UK in a single operation rather than separate operations for different countries.
- Integration of data and metadata to support understanding and effective use
Wednesday 26th March 2014, 1-4pm
RSS Glasgow/Edinburgh/BIS joint event: Scotland's Referendum: Statistical Perspectives
***PLEASE NOTE THE CHANGE OF VENUE: *** Paterson's Land, Holyrood Road, EH8 8AQ Edinburgh
Six months in advance of the referendum on Scottish independence, the Glasgow and Edinburgh Local Groups and the Business and Industrial Section of the RSS will hold an event in which four speakers will offer their perspectives on statistical aspects of the referendum and its potential consequences:
Tuesday 15th April 2014, 6-7pm
The Early History of Bayesian Statistics
Tom Leonard (Retired from Universities of Wisconsin-Madison and Edinburgh)
ICMS, 15 South College Street, Edinburgh EH8 9AA
Tea & coffee available from 5:30pm
Thomas Bayes studied at the University of Edinburgh between 1719 and about 1722. Early contributors to the philosophies and methodologies which were to influence Bayesian Statistics also included William of Ockham, Daniel Bernoulli, Richard Price, Pierre-Simon de Laplace and, during the Scottish Enlightenment, David Hume and Adam Smith.
The speaker will discuss the subsequent history of the Bayesian paradigm until its mid-twentieth century resurgence, which was initiated by the work at Bletchey Park during the Second World War by Alan Turing and Jack Good, and by Sir Harold Jeffreys' novel research at Cambridge. Jeremy Bentham, Augustus De Morgan and Sir Francis Galton, all of University College London, played significant roles in the nineteenth century history.
Monday 19 May 2014, 6-7pm
A joint event with University of Edinburgh Business School on:
Perceptions of the Weight of Statistical Evidence among Scientists
Speaker: Richard D. Morey
(Assistant Professor, Psychometrics and Statistics, University of Groningen, Netherlands)
Venue: Conference Room and Roof Terrace - University of Edinburgh Business School (29 Buccleuch Place, EH8 9JS)
Social: Tea & coffee available from 5:30pm, and a network drink reception is from 7pm afterwards.
In science it is critical that claims be backed by evidence. Within the social sciences, evidence is often statistical in nature. Statistical outcomes such as p values or confidence intervals are frequently used to communicate results, and are often described using terms like "strong evidence", "weak evidence", and even "no evidence" which seems to imply that scientists believe that the weight of statistical evidence can be quantified using commonly-reported statistics. We present the outcomes of a questionnaire of scientists in which scientists were asked to evaluate the weight of statistical evidence in fictitious studies reported with commonly-used inferential statistics. Perhaps surprisingly, we found that many scientists deny that the strength of evidence is quantifiable given these statistics. Furthermore, scientists indicating that the weight of evidence was quantifiable were enormously variable in their estimates. The unconstrained nature of these epistemic judgements has potentially severe implications for science.
Thursday 16 October 2014, 6 - 7:45pm
A joint event with the University of Edinburgh Business School on "Big Data from Two Perspectives":
Venue: The University of Edinburgh Business School
Two speakers will offer their perspectives on:
Speaker one: Prof Marian Scott.
Affiliation: School of Maths and Statistics, University of Glasgow.
Title: Are there New Challenges in Environmetrics Arising from "Big Data?"
In this presentation, I will consider current and some future statistical challenges in describing, modelling and forecasting the state of the environment from a statistical perspective. Examples will be drawn from air, water and soil.
Speaker two:Prof Mark Keane.
Affiliation: UCD & Insight: Centre for Data Analytics, Chair of Computer Science,School of Computer Science & Informatics, University College Dublin.
Title: (Big) Data Analytics: From Word Counts to Population Opinions.
Behind the hype surrounding the commercial promise of Big Data, a very real change has occurred in the type of research being done in Data Analytics. Though many of the statistical techniques used are not particularly surprising or new, often the way they are used, combined with advances in their computational tractability, make them seem quite novel. A second key novelty in this research development is the shift to considering "non-traditional", unstructured data sources such as text, tweets, web-searches, facebook ratings and so on. This talk will review some current interesting examples of Data Analytics, with a specific emphasis on text analytics.