Smartphone Notifications Dataset available for download

We have released a rich dataset containing 176,195 logged notifications received on the Android devices of 14 users, in the period between 2018-05-07 and 2018-06-05. The dataset accompanies our publication Komninos, A., Frengkou E., & Garofalakis J. (2018). Predicting User Responsiveness to Smartphone Notifications for Edge Computing. 2018 European Conference on Ambient Intelligence (AmI-18). Larnaca, Cyprus, Springer. .

Please free to use for your own research, citing our work if you found the dataset useful.

New publication: Predicting User Responsiveness to Smartphone Notifications for Edge Computing

Edge computing requires the addressing of several challenges in terms of privacy, complexity, bandwidth and battery life. While in the past attempts have been made to predict users’ responsiveness to smartphone notifications, we show that this is possible with a minimal number of just three features synthesized from non-sensor based data. Our approach demonstrates that it is possible to classify user attentiveness to notifications with good accuracy, and predict response time to any type of notification within a margin of 1 minute, without the need for personalized modelling.

Komninos, A., Frengkou E., & Garofalakis J. (2018).  Predicting User Responsiveness to Smartphone Notifications for Edge Computing. 2018 European Conference on Ambient Intelligence (AmI-18). Larnaca, Cyprus, Springer.

New publication: A glimpse of mobile text entry errors and corrective behaviour in the wild

Research in mobile text entry has long focused on speed and input errors during lab studies. However, little is known about how input errors emerge in real- world situations or how users deal with these. We present findings from an in-the-wild study of everyday text entry and discuss their implications for future studies.

Komninos, A., Dunlop M., Katsaris K., & Garofalakis J. (2018).  A glimpse of mobile text entry errors and corrective behaviour in the wild. Extended Abstracts, Mobile HCI'18. Barcelona, Spain, ACM. DOI:10.1145/3236112.3236143

Judge at CrowdHackathon SmartCity 2 Citylab event (Patras)

I was invited as a scientific expert to be on the judges panel for the Patras Citylab (prep event) for the City Challenge Crowdhackathon #smartcity 2 event. A total of 9 ideas were presented by teams and solo applicants, and we selected the best 3 to move forward to the main event with all their expenses covered (Athens, 28 June - 1st July).

Congratulations to all the winners and good luck to every team that decides to move on to the main event!

New publication: Measuring Inviscid Text Entry Using Image Description Tasks

We argue that measuring the Inviscid text entry rate requires new evaluation methods that support freeform text entry and that are not based on the traditional transcription/copy tasks. In this position paper we propose use of image description tasks and share some of our experiences of using this new language agnostic task type for free form text entry

Dunlop, M., Nicol E., Komninos A., Dona P., & Durga N. (2016).  Measuring Inviscid Text Entry Using Image Description Tasks. ACM CHI’16 Workshop on Inviscid Text Entry and Beyond. San Jose, USA, ACM.

Pages