Looking forward by looking back

Janus

As the Small Data Forum progresses through its early teenage years – our latest podcast is episode 14 already – regular co-hosts Thomas Stoeckle, Neville Hobson, and Sam Knowles are taking the opportunity to look forward by looking back.

Patients of our own medicine, you might say, we’re using the year end and what we’ve observed and learned in 2017 to enter the predictive analytics business.

We take our inspiration from Janus, the Roman god of beginnings, transitions, and time, after whom January is named. A sculpture of Janus appears at the top of this blog, from the Vatican Museum.

Continue reading “Looking forward by looking back”

04: Brexit, Trump and the challenge of better forecasting with better data

Our latest Small Data Forum podcast focuses on recent surprise political outcomes and the mechanisms at play: the psychology of predictions, the new phenomenon of fake news, echo chamber effects and the way data was analysed and interpreted.

Recent expressions of democratic political will – the UK referendum on EU membership, the US presidential election – have surprised most observers and commentators.

Both outcomes, ie Brexit and Trump, were not what most of the polling data indicated. This episode of the Small Data Forum is asking whether we should and could have seen this coming.


Continue reading “04: Brexit, Trump and the challenge of better forecasting with better data”

02: The role of data in Regrexit

Episode 2 of our Small Data Forum podcast focuses on the role and use of data in the campaigning for the EU referendum in the UK.

Neville Hobson, Sam Knowles and Thomas Stoeckle discuss the outcome of Brexit and Regrexit (the so-called buyers’ remorse of some leave voters, following an immediate change in the argument and presentation of facts, as well as data evidence of negative effects on currency, share prices etc.).

Ranging from failed predictions and misinformation, to the need for more learning and education related to data, to the use of data visualization and linguistic analyses and Bayesian statistics to segment audiences and identify archetypes, the conversation focuses on how data can and must become the solution rather than the problem.


Continue reading “02: The role of data in Regrexit”