Get control of the swarm. See the swarm moving in advance before something can happen
What is the program? Social Media? Why What is Social Media Monitoring? Who does it? How to do? Results? Recap
Social is not new. Where we used to listen on a scale of platforms offline, we now use technology to receive information online on a wider range of platforms
People having conversations online In the early days we were not social either
Snapshot of playing field. This playing field changes every day, week. Platforms come and go, Hyves, Myspace, Google+, etc.
This is a view of a tweet with all data we can collect. We can analyse the data cross platform and within different data ranges. The graphics and charts give a detailed view
Preview of results
Use multiple channels to get an overview of all buzz online
This is an example of the control room at Tilburg fair
This is an example of the control room at Decibel. We were close to the operational team for security and medics
Where to begin? When to begin? Start some days in advance to get to know the visitors and fans. In these days you see that most of the conversations are before the game, till an hour before the game starts
If the atmosphere is ok, all goes well. When conversations show that more negative messages are placed you have to monitor closely.
This is a list of all people at the match #feyaja
This trendline shows us that people having conversations and that possible incidents can happen. Closer to match day, the number of conversation drop and conversations about the game itself are rising.
Here you see that conversations take place before on and after the match. There is a slight difference between in advance and after.
By monitoring sentiment as in positive and negative you get a detailed view of the temperature in the crowd. With cross analysis of sentiment you can monitor specific groups of fans and get insights about their conversations. We can also signal if possible incidents can happen. By monitoring sportevents we read messages right at the moment when they happen. By using trendlines of volume you are in the possibility to predict some events.
Examples tweets and instagram posts
What to do on matchday? With a team of two persons we read 13.000 messages around the match Feyenoord Ajax. We added metadata as labels. These labeled messages automatically were published in reports. These reports were sent to security and police to give a detailed view of the atmosphere.
Preview of tagged results
Preview of reports. Depending on the messages and notifications action are needed. For instance: people parking bicycles alongside the railway track because lack of bicycle parkingspace. The result is a dangerous situation for people and trains. With online monitoring we get an early view of these situations. Because people tweet, post messages and share images we notify these signals before the train machinist knows. So monitoring helps also to give better service. When you notice such an incident you can respond and take action to prevent this situation the next match
On matchday people still are having conversations. In the stadium the volume is less than before the match. Possible explanations are no signal and no wifi, attention for the match or batteries of mobile phones are empty. This is what we see at several events near the end of the day. Negative sentiment does not always mean aggression. The negative sentiment is part of the result of the match.
What do we monitor? Different search queries: just words and sentences with knowledge from the days before the match. Geotagged posts within the area of the stadium, but also on the route to the stadium Specific persons and possible targets
Some results: We do not only listen to regular Dutch verbs, we also learn about specific football and curses. People also tweet with spelling errors. You have to update the search queries continuosly
Monitoring is also used for proof. This flag from Ajax youth was placed at the Erasmusbridge in Rotterdam. The flag was already removed before most of the crowd noticed. A social media monitor collects all data and saves it. If someone posts a message and removed it after we have it still in our archives. So online messages can be used for confirmation and detection
2014 09 25 Sport and its supporters: keep in touch with the crowd
Keep in touch with the crowd
Sport and its supporters
#SMWRdam | 25 september 2014 Martijn Rijk | +31 6 29 437 111 | @martijnrijk
Start monitoring some days in advance
use of language
most active persons
Ben de Graaf (30)
hashtags (#feyaja and #klassieker)
location, traffic and facilities, inc. gps tagged
images (photo en video)
posts from visitors
Specific people like Kenneth Vermeer, Fred Rutten,
‘war language’ and footballjargon
most curse before and after match
most video/photo content before the match
traffic dominant 2,5 hours before match
Work as a team with security, medics and club
communicate via hangout/whatsapp/porto
use multiple tools to monitor
updating search queries is an ongoing process
listening is a start
interact on incidents and questions