Dr. sc. nat. Kiran Zahra
Dr. Kiran Zahra is a visiting researcher at GiScience Research Group of the Institute of Geography in Heidelberg. She is a fellow of the Swiss National Science Foundation (SNSF) and was awarded for a PostDoc Mobilty Fellowship.
- Information extraction
- Unstructured social media data
- Mass emergency events
- Opinion mining and sentiment analysis
- Machine learning
My research is about making sense of unstructured social media posts published during mass emergency events. Social media data contains the most updated information about an ongoing event, however, due to its unstructured and informal nature of communication, it is mixed with a lot of noise. I use machine learning algorithms and other computational methods to extract only relevant and credible information out of a bulk of data automatically. My current projects are to analyse social media posts for heat stress-related messages to develop a schema that translates relative linguistic terms used to describe the phenomena into absolute temperature values to generate heat maps for a region. Moreover, I am also interested in analysing small-scale events that are triggered by mass emergency events. As these small-scale events have a limited vocabulary, therefore, they often go unnoticed in the social media stream. Whereas, these events have an important impact on a local community and contain useful information for emergency responders.
- Zahra, Kiran and Purves, Ross (2017). Analysing tweets describing during natural disasters in Europe and Asia. VGI-Analytics, AGILE, May 9-12 2017. Online, http://www.cs.nuim.ie/~pmooney/VGI-Analytics2017/materials/Zahra_Purves_AGILE_2017.pdf
- Zahra, Kiran, Ostermann, Frank, & Purves, Ross (2017). Geographic variability of Twitter usage characteristics during disaster events, Geo-spatial Information Science, 20:3, 231-240, DOI: 10.1080/10095020.2017.1371903
- Zahra, Kiran (2018). Natural disaster database design and development for Himalaya using social media. In Proceedings of the 4th AGILE PhD School (eds. Lex Comber and Nick Malleson), Leeds, UK, October 30 - November 02, 2017, CEUR-WS.org, online, http://ceur-ws.org/Vol-2088/paper9.pdf
- Zahra, Kiran, Imran, Muhammad, & Ostermann, Frank, (2018). Understanding eyewitness reports on Twitter during disasters. ISCRAM 2018 Conference Proceedings. 2018.
- Zahra, Kiran, Imran, M., Ostermann, F. O. (2020). Automatic identification of eyewitness messages on Twitter during disasters. Information Processing and Management, 57(1),102107
- 04.2021 – current Postdoctoral researcher, Geoinformatics, University of Heidelberg
- 09.2015 – 03.2021 PhD researcher, Geocomputation, Department of Geography, University of Zurich, Switzerland
- 10.2007 – 12.2009 Master of Science in Geographic Information Science and Remote Sensing, National University of Sciences and Technology, Pakistan
Media coverage of my research
My PhD research in 3 minutes