Yunus Emre InstituteAs big data seeps into everything we do, the Yunus Emre Institute in London kicked off the 2021 iteration of its Young Scholar Seminar Series on Tuesday 20th April 2021 with a highly relevant discussion on data analytics:
“Forecasting Turkey’s 2017 Presidential Referendum Using Big Data”
With the help of moderator Aysegul Chouseinoglu, the speaker, Abdullah Cemil Akçam, talked the audience through how the team had predicted the outcome of the Turkish referendum from amongst misleading polls.
Social media platforms have become key mediums for conducting political discourse and voting preferences. As Akçam set out in his talk, Twitter has emerged as the main discussion platform for this exchange of views as well as political campaigns, making it a potential source of information to be analysed and explored to gauge voting intentions. With Twitter’s widespread use as a medium of communication by a high number of active users, the sheer volume and variety of the created data requires new technologies, such as big data, to be employed.
Big data, a field that extracts useful information from data that is huge in size and growing exponentially, became a buzzword recently, and attracted a lot of attention across organisations. Akçam explained that Twitter became the core source of this research as it fulfilled the so-called “3 Vs” of big data: volume (10 million tweets per day), velocity (100+ tweets per second on peak hours) and variety (fast-changing data). Making highly technical and complicated processes and information easy to understand, Akçam talked through how his model worked. For example, he showed the fascinating raw format of a tweet and how data analysts made sense of it.
Continuing the talk to touch on the methods of analysis he used and why, Akçam delved into the various mathematical, technological, and surprisingly human elements of data analysis. He noted the potential shortfalls in the project but explained how they were overcome to develop a model that could detect voter sentiment with a high accuracy rate of 94.6%!
Given the inability of polls to always approximate voting intentions over the past years, this discussion on methods of predicting referendum outcomes was a fantastic introduction to data science and analytics for the audience. The seminar ended with many questions from the audience that inevitably arise when discussing hot topic issues.
Akçam studied Computer Engineering at Middle East Technical University in Turkey and attained an MBA at Bilkent University while working at various organisations. In 2017, Akçam was awarded the Chevening Scholarship and studied MSc in Business Analytics at Queen Mary University of London. He has currently returned to Middle East Technical University as a PhD candidate.
This talk was part of the Young Scholar Seminar Series, also known as Genç Akademisyenler, an international and interdisciplinary platform for knowledge exchange established in 2017. Each seminar features an outstanding postgraduate student who is studying a global or Turkey-related topic presenting their research to the wider public. The long-running series, is an opportunity for all those interested, expert or those simply curious, to engage and interact with academic leaders of tomorrow.
The upcoming weeks will see further scholars take the virtual stage and discuss their research areas. These scholars include:
-
-
- Hajar Saihi: From immune cells to big data, making it all make sense
- Abdulbaki Turkmenoglu: Copyright Law in the Age of Artificial Intelligence – Redifining “Author”
- Hande Alptekin: Climate Change: Critical Role of Batteries
-
Recordings of the talk can be viewed on Yunus Emre Institute in London’s YouTube channel, Yeelondra.
For updates on future Young Scholars please visit the Institute’s website, https://yeelondon.org.uk/, or follow their social medias, @yeelondra.