Twitter experiment at Royal Meteorological Society Conference

Michelle Spruce recently attended the Royal Meteorological Society (RMetS) Student & Early Career Researcher conference at the University of Birmingham on 4/5 July 2019.

As well as opening the conference by presenting her research on the social sensing of extreme weather events, Michelle also encouraged conference attendees to use Twitter during the conference in a social sensing experiment to understand the impact of ‘tweeting’ during an academic conference.

Over the 2 days of the conference attendees tweeted news and updates using the conference hashtag #RMetSStudents. By lunchtime on the second day of the conference with just 162 tweets Michelle was able to demonstrate the wider impact of these tweets:

By the end of the conference, 203 tweets including this hashtag were generated, from 44 users in 6 countries and 13 cities. While a seemingly small amount of data, by the end of the conference these tweets generated a potential reach of over 32,000 Twitter users and over 500,000 impressions (individual views of these tweets). This simple experiment demonstrated the power of using Twitter as a source of information even for small scale events such as this.

Recent work: Social Sensing of named storms in the UK and Ireland

Michelle Spruce recently presented some of her research on social sensing of extreme weather events at the Royal Meteorological Society’s public ‘Weatherlive’ conference in November 2018.  She also discusses her work on the ‘Paul Hudson Weather Show’ which broadcasts on BBC Radio Leeds, York, Sheffield, Humberside, Lincolnshire and online in early December.

Her research uses Twitter data collected during the period of the 2017/2018 UK storm season (Autumn 2017 to Spring 2018).  Building on the work already done in the research group on the social sensing of floods and hayfever/pollen, this study aims to determine the social impacts of named storms in the UK and Ireland.  Storms are named when they are forecast to cause moderate to severe impacts. To improve data quality, tweets are filtered for relevance to the named storm event using simple text filters and a more complex Naïve Bayes machine learning approach.  After removing irrelevant tweets, we find peaks in Twitter activity which correspond to the time period of the storm.  Using the filtered data, we also calculate a sentiment score (how positive or negative the tweet text is) over time.  We find tweets becoming less positive during and in the hours after the peak of stormy weather.  Categorisation of tweet content during the storm period also finds more than a quarter of tweets are grouped within the ‘humour category’, and a further fifth of tweets reporting on damage or disruption.  Using the findings from this research will help to better inform impact based weather forecasting and also provides an additional forecast validation tool.

Michelle is hoping to submit her findings to a relevant journal for publication shortly.

Storm Brian tweet density during the storm period 21st October 2017
Sentiment polarity score for Ex-Hurricane Ophelia tweets vs Tweet count