data

Our first Virtual CDAS Partner Engagement event

by

On the 24th September 2021, the Centre for Data Analytics and Society held its annual partner event online for the first time. Attended by academics from across the CDT institutions and representatives from partner organisations, the event proved a great opportunity for students to get together and share their work, catching up or in some cases meeting for the first time!

The event was planned and delivered by a selection of students from multiple cohorts and home institutions, working together under the guidance of CDT management to plan an engaging and enjoyable virtual conference. Students moving into Year 3 and 4 were split into theme groups and delivered lightning talks on aspects of their PhD projects, while students moving into Year 2 presented posters in a virtual networking space. You can view the posters at the bottom of this article.

Students gather in a virtual poster session in GatherTown
Instead of presenting in a physical space, students could view and present their research posters in a virtual room.

Representatives from the Department of Work & Pensions, Esri, Sheffield City Region, HERE Technologies and Geolytix joined the event to deliver presentations on their work and to participate in a panel discussing how data science is used in their field, and providing tips for PhD students looking to move into commercial or public sector work after graduating.

The event closed with a pub quiz in GatherTown, hosted by our student delivery group. We’re keeping our fingers crossed that next year, we’ll be able to celebrate another successful event in a real pub together!

Research posters presented by Year 2 students

Using Content And Sentiment Analysis To Understand Populism In Donald Trump’s Speeches- Ruth Neville

Text mining & sentiment analysis of Boris Johnson’s lockdown announcements- Nastazja Laskowski

Spanish house prices, What factors are important- Gladys Kenyon

Predicting strokes- Ron Bar-Ad

Predicting Outcomes of Apprenticeships Delivered within Sheffield City Region using Machine Learning- Shivani Sickotra

Police Twitter presence in the period after George Floyd’s death- Erica Kane

How Reliably Can Vivacity Cameras Count People- Mushtahid Salam

Green Homes Grant- Cameron Ward

Food choice patterns of primary school children -Ines Francois

Dimensionality Reduction for Identifying Valuable Airline Customers- Amy Jungmin Seo

A machine learning approach to understanding factors- Ellie Marfleet

 

End of year update from our fourth CDAS cohort

by

Hello again from the fourth cohort of the CDT Data Analytics & Society (CDAS) program. Today we will be sharing our experience of our journey through the second semester of our time here at the CDAS (in case you missed the post about the first semester, it’s here!). We (Javiera and Mushtahid) have structured this blog post in basically two parts. In the first part, each of us will describe our favourite part(s) of this semester, while in the second part we will describe our journey through this semester from a more global perspective.

Javiera: “My favourite part of the second semester was the internship experience with my sponsor company, the Office for National Statistics (ONS). I worked with data from the Opinions and Lifestyle Survey (OPN) which was created to understand the impact of the coronavirus (COVID-19) pandemic on British society. I collaborated with the social care team to study the impacts of COVID-19 on unpaid carers during the month of April 2021. I used the programming language R to analyse the data: the results were the basis for a publication released on the ONS website earlier this month. It was rewarding to have been involved in the project and seen it evolve from beginning to end. The internship was also an excellent way to meet and engage with the social care team at the ONS. It was interesting to learn about how a data provider like the ONS works. And although the internship was completed online, the experience felt very social. It was extremely motivating to work as part of a team after a year of little interaction with peers.”

Mushtahid: “My favourite part of the second semester was most probably when I learned about neural networks and their vast array of applications in solving real-life problems. Neural networks could play a huge role in my PhD – which at this stage is mostly focused on measuring physical activity and other wellbeing behaviours of people in urban spaces using digital tools. Therefore, I could certainly see myself using neural networks to analyse video recordings to detect people and recognise their different physical activities in urban areas. In fact, my internship this semester involved working with a dataset provided by a tech company that specialises in building real-time object detection systems via deep learning to capture different road users such as pedestrians and cyclists. The internship was certainly an enriching experience as it involved me collaborating with my industry partner Buro Happold and the tech company. Similar to Javiera’s experience, my internship felt social despite it being done entirely online.”

During this semester, we also learned about the importance and application of metadata, privacy, confidentiality and anonymisation in today’s data-driven world through the module “Understanding Data and its Environment” taught by Prof Mark Elliot and Dr Nuno Pinto at the University of Manchester. In the same module, we were taught about data pre-processing by Dr Pradyumn Shukla and were tasked to pre-process a dataset and subsequently build a predictive model to solve a real-life business problem – which enabled us to gain valuable insight on how to properly prepare datasets so that they are ready for analysis.

Currently, we are going through our final module for this semester – “Social Analytics & Visualisation” taught by Dr Mark Taylor, Prof Paul Clough and Dr Griffith Rees at the University of Sheffield. In this module, we are being taught how to do data visualisation, textual analysis, machine learning and social network analysis using R. Although it is being conducted online (similar to the other modules) because of the pandemic, we are grateful to have an excellent academic staff and a fantastic CDAS team to ensure that our journey is as smooth as possible. For example, Hayley Irving has been conducting regular “Shut Up and Write!” sessions to help us focus on our studies while working from home, and Claudia Rogers has been conducting regular drop-in sessions where students can talk to her about anything they are struggling with. Moreover, we have also had a buddy system set up through which senior CDAS students are guiding us through this journey. They have also written helpful blog posts such as PhDing in a pandemic: a guide on surviving. We want to thank them for their continued support.

Article by Javiera Leemhuis Arnes (University of Sheffield) & Md Mushtahid Salam (University of Manchester)

Introducing our fourth student cohort

by

Hello! We are the fourth cohort of the CDT Data Analytics & Society program. Navigating our first semester on this integrated masters and PhD course has been a very different experience, but we have adapted successfully to online teaching. We have found it challenging as a cohort to connect, considering many of us have never even met in person!

Our first module was named Programming for Social Scientists with The University of Leeds. This was a two-week intensive module taught by Andy Turner who brought our group together, as well as equipping us with the necessary Python programming skills which our PhD projects will require. Even though this two-week intensive course tested our limits, we can all agree we gained valuable skills including introduction to Agent Based Modelling, creating our own website and learning how to use GitHub. We began our social research modules taught by our home universities – either Leeds, Sheffield, Liverpool or Manchester. Since our undergraduate backgrounds varied from Mathematics to Psychology these modules introduced us to new ways of thinking, preparing us for undertaking our own research in the coming years.

We are excited to continue our journey of learning with the upcoming Data Science Studio module delivered by Dr. Daniel Arribas-Bel at The University of Liverpool. Although, sadly we cannot live and learn in Liverpool as many of us were expecting, we are still looking forward to supporting each other online.

We would also like to give a huge thank you to all the team members who have supported us and helped us to integrate into the Data Analytics & Society program. We are all looking forward to starting semester 2 in January!

Article by Cameron Ward (University of Liverpool) & Shivani Sickotra (University of Sheffield)

 

Introducing Dr Henri Kauhanen, Postdoctoral Fellow

by

Dr Henri Kauhanen is ESRC Postdoctoral Fellow with the Data Analytics & Society CDT from October 2018 to September 2019. Affiliated with the division of Linguistics and English Language at the University of Manchester, Henri works on mathematical and computational models of the population dynamics of language, looking for explanations of universal factors of linguistic variation and change that recur from one language to another. Originally trained as a cognitive scientist, Henri received his PhD in linguistics from the University of Manchester in April 2018, supervised by linguists Ricardo Bermúdez-Otero and George Walkden (now at the University of Konstanz) and theoretical physicist Tobias Galla.

 

For his one-year ESRC Postdoctoral Fellowship, Henri is taking a data-driven approach, concentrating on a focal topic in language dynamics but one that has so far received surprisingly little attention. This is the question of what basic rates different features of human language evolve at, and what is to be made of cases where several changes are governed by identical rates of change. To this end, he is writing computer software that aids in fitting the predictions of different models of linguistic change to empirical data, as well as conducting Monte Carlo power analyses of existing methods for model selection, utilising Manchester’s HTCondor high-throughput computing framework. The software will be released as open source packages for the R statistical computing environment by the end of the fellowship.

 

In addition to work on implementing computer code and writing research articles, the project includes a significant networking and skills development component. Henri attended the 2019 Complex Systems Summer School run by the Santa Fe Institute [link: http://www.santafe.edu] in Santa Fe, New Mexico in June and July, attending lectures on complexity, chaos, networks, nonlinear dynamics and related topics, but also working on group projects with physicists, biologists and social scientists. In August, Henri is organising a symposium on language change at Manchester [link: http://rusesymposium.org.uk]; alongside regular talks, the two-day long event will feature three keynotes by eminent scholars in the field, focusing on resolving some of the often considerable tension between computational modelling and empirical, data-oriented work.

 

In the future, Henri is planning to continue working on computational models of language change, aiming in particular to increase the realism of currently available models. In October, he is moving to Germany to take up a second postdoc at the Zukunftskolleg [link: http://www.uni-konstanz.de/zukunftskolleg], an Institute for Advanced Study for Junior Researchers at the University of Konstanz.

 

To find out more about Henri’s research, visit his website at http://henr.in [link: http://henr.in].