�ݺ�ߣ

Editor's Notes

• #2: Good morning My name My origin Title of talk The Social journeys project – exploring how social media updates can be combined with existing (open) datasets to further enhance real-time passenger information. Team – design researcher, computer scientists and transport researchers Today I am going to show you how to combine social media data with fieldwork data to help understand how a company engages with their customers through this channel. Background – emergent use of social media in public transport and the company where the study was conducted Method – the challenges of dealing with different types of fieldwork data alongside social media data Results - will describe some of the emerging themes from the study with the operators use of social media and the implications this has on their communication channel
• #3: Social media is being seen as an increasingly important channel for communication For journey planning, it has been used to predict the demands of passengers in order to ensure the service levels match requirements of passengers For transport policy, social media data has been mined to support the needs of transport system planners and policy makers It has been used in minor travel disruption to help provide travel information advice if there is an issue which has caused problems in the transport network. It has also been used during major disruption such as a natural disaster to help guide the transport operator’s response such as cancelling or rescheduling services On the right is an example of a minor disruption reported via social media. This is from TfL which is Transport for London report from Oxford Circus which is one of the busiest sections of London. Created quite a storm on social media – as you can see with the number of retweets and favourites.
• #4: The organisation where the study was focussed was FirstGroup. They are the world’s largest - In the states they run some of the school bus services and the greyhound which are scheduled intercity coach transportation service. UK bus – 18 subsidiary companies 3 Scottish subsidiaries were selected to provide a comparison of the social media practice in these different subsidiaries.
• #5: FirstGroup in Scotland They had different characteristics in terms of scale and types of population served First Aberdeen has an urban population with 160 buses operating First Glasgow had a much larger operation with 1000 buses operating in a mostly urban setting In between this in terms of scale with 420, was First South East and Central who had a rural population feeding into a larger urban location with Edinburgh. Next I am going to discuss the methods that were deployed for our study.
• #6: Moving onto methodology and how we went about the study. This required us to deal with different types of data. There was online social media data gathered through Twitter with a content analysis applied to help categorise and define the content of tweets. This data was extended through fieldwork at the operators depots to help understand their practice for posting social media generating 8 hours of audio and 2 weeks of observations A co-reflection activity taking the analysed social media material and presenting this back to the participants for further reflection. Speaking to the theme of this session, there were an emerging practices that was deployed to help shape the analysis: dealing with Multiple types of data – the content analysis helped to frame the enquiry for the fieldwork activity. This then flipped back again so that following the fieldwork, we could go back to the feeds and see how our understanding of their practice was then playing out through the feeds themselves. We’ll now go into a little more depth for the process for each of these starting with the content analysis
• #7: When processing social media data there were some specific conditions that were applied for doing this: We had to make sure to capture the conversations between the operator and passenger. Where there were messages addressed to the operator, and they operator had responded, it was important to follow the threads both up and down to identify how the conversation had started and be subsequently resolved. To help in collecting the conversations, the Twitter API was used to build a Twitter Monitoring Infrastructure. This crawler the accounts and kept an eye on the real-time conversations that were taking place. The collection of the social media data started before the fieldwork activity and continued whilst the fieldwork was underway. TMI is open source and available on GitHub, if anyone is interested – please come and chat to me.
• #8: The scale of the data was confined to the tweets that were flowing through the operator’s accounts. A challenge when dealing with social media data is the shear scale of material available. Keeping it focussed on these operator accounts gave us a good strategy to manage this. This gave us 1672 tweets over a four week period in Jan and Feb this year. There were no major events events causing major disruption taking place during this time so it was a normal amount although given the time of year, the weather conditions were sometimes challenging. When classifying tweets, some messages fell into multiple categories so this was accounted for. The conversations across multiple tweets could also be categorised so these were also assigned meta categories. The schema categories used were promotion, feedback, travel information and response. So drilling down into promotion – this included the morning greeting ritual when the operator signed on, timed press releases and other marketing material, chat with passengers and happy staff
• #9: The people who were engaged with were the Twitter operators. These individuals had different roles within the company. Most of the people behind the feeds were part of the marketing team. However, a street inspector works on the buses checking people have tickets. A network manager who co-ordinates the service was also operating the feeds. This was supplemented with fieldwork observations in their offices and the operator depots to understand how the staff co-ordinated themselves and worked with other parts of the organisation. These people were spoken to as they were mediating the contact between the company and their customers so had direct experience of the daily issues that are encountered. There were a range of questions that we went into the field with.
• #10: Fieldwork questions were shaped by the previous content analysis. They were also refined into an interview protocol following initial conversations with the marketing manager at one of the field locations. Strategy - Application of social media as communication channel Crafting of Content - Protocol for posting material and responding to passenger tweets Flow of real-time information - Use of social media as an input for updates on unplanned disruption Internal perception - Employees perspective including drivers through to managers of social media and the affects on customer service
• #11: There are much more results in the paper so have concentrated on the method to fit with this session but please do read the paper and seek me out if you would like to discuss any details further. persistent conversation being engaged in with the extension of customer relationship beyond duration of journey through the use of twitter. If a passenger is taking a regular service where there are different drivers each time, twitter has the same set of operators so is able to provide a familiar virtual face. There were different approaches for the provision of real-time information. The scale of the operation and availability of information affected whether the operator was reactive or proactive about pushing updates concerning disruption to the service. Personas were used as a means to relate to passengers and these was carefully managed through the accounts. There were different approaches to the use of persons in each location which affected how the operators and passenger related to one and other. For example, in one location, there was an individual persona for each of the 4 twitter operators so this gave them greater ownership over the personas. Whereas in another, there was a single persona with 3 people behind it. A collapsing of roles taking place where those managing the twitter feed came from a marketing background but this traditional role was being supplemented with providing real-time travel information. There was also a distinct customer service aspect to there role as they help passengers on their journeys and channel any enquiries to the appropriate parts of the organisation.
• #12: Visualising Twitter data Fed back some of the earlier content analysis results to participants as a starting point to ask them to reflect on their own feeds and relate them to others. Comparison The differences between strategy of each subsidiary quickly made apparent from visualisation. Those on the right from top to bottom show the Aberdeen, Glasgow and South East and Central visualisations. The dominant personas, frequency of retweeting and strategy to take complaints offline were distinguishable between these accounts
• #13: When focussing on social media data, we kept our focus on a few specific, comparable accounts and practices around these. The content analysis worked really well as preparation for the fieldwork activity. The strength of it was the twitter monitoring infrastructure we had built could be deployed whilst the fieldwork was underway. This meant that we were collecting data back at our base whilst being able to interview and observe the operators as they were managing the feeds so this gave us multiple perspectives on the same things happening. Getting initial access to the companies was a challenge but it took time to build up this relationship. The use of Twitter for providing travel information could be greatly altered during larger events with a higher volume of data being produced and different conditions for managing them so this would provide further area of interest. To generalise the method into different business contexts, the method of combining fieldwork data and social media data as we have done provided considerably more understanding of the role social media plays internally within an organisation for supporting customer experience than either could alone.