Project description

Migration Trail is a mapped data visualisation following the journeys of two migrants, setting off from north Africa and Turkey, travelling to and through Europe in real time, over ten days. For ethical reasons, this is a reconstruction, but is based on true stories. The voices of the two migrants – a Nigerian man and a Syrian woman – have been written as instant message feeds by Nigerian author Elnathan John and Lebanese scriptwriter Nadia Asfour. The audience follows the story on the website, with the message feeds also available in Facebook Messenger, with people receiving them on their phone, as the story unfolds, wherever they are. The real time telling of the story makes it urgent and immediate, while the use of maps and data brings a new approach to an issue that many people feel they already know well, though they typically don’t. We wanted to create a more informed discussion about irregular migration and to show that maps and data can tell a compelling story. A podcast series follows a number of real-life migrants’ and explores the wider issues.The map changes colour over the course of the day and night, while data embedded in it gives a sense of the conditions of the journey: the availability of wifi, wind flows, shipping traffic, flights over Europe. The sound design serves as an emotional barometer for the piece. Zooming in, you see personal data about the migrant character, such as the distance they can continue to travel before their phone battery dies. Zooming out, you see the data that puts their individual experience into a wider context – statistics about where people from a given country apply for a asylum and where that request is more likely to be granted. The map shows the links between individual experience and the social and political factors which shape it. The journey serves as a framework for understanding how the various stories that people may have heard on this topic – of Libyan people smugglers, Hungarian border closures and camp evictions in Calais – are in fact part of the same, wider story.Our target audience was European: the UK and the Netherlands, in particular. The instant message feeds were available in English, Nigerian Pidgin, Dutch and Arabic, to make the experience accessible for UK and Dutch audiences, as well as people in key countries of refugee origin. While we aimed for a broad general audience, in our outreach we first aimed to reach either those with an interest in data visualisation and documentary, or those with an interest in the subject of migration. Migration Trail launched in November 2017. In the first ten days 4,390 people visited the site, with 15% visiting multiple times. They spent an average of 10.46 minutes on site. We retained 67% of our audience over the ten days using the Facebook Messenger bots. There was no monetisation plan for the project. It was funded by a WIRED/the Space Creative Innovation Fellowship and grants from the Creative Industries Fund NL, Netherlands Film Fund, Dutch Media Fund and Arts Council England.

What makes this project innovative?

This project solves a number of challenges in storytelling projects with maps to date, by adding further structure to make them effective for storytelling. Where previous mapping platforms allowed journalists to add text, video and images to maps, they have typically lacked any mechanism to tell viewers what to look at and in what order, so that it is impossible to direct a narrative. In this platform viewers' attention is directed by using the structure of an unfolding journey. This gives a clear viewing sequence to the information on the map. Telling the story in real time paces the release of information, with new statistical maps added day by day as they add relevant context. The real-time element also makes the migrants' stories urgent and immediate for the audience.The zoom mechanism adds further structure. Zooming in, you see personal data about the migrant character, such as the distance they can continue to travel before their phone battery dies. Zooming out, you see the data that puts their individual experience into a wider context – statistics about where people from a given country apply for a asylum and where that request is more likely to be granted. The map shows the links between individual experience and the social and political factors which shape it, helping to give a complete picture of an issue. We also experimented with different ways to display datasets – the missing migrants data is usually displayed on a map with different sized circles to show the number of deaths in different locations. We disaggregated this data, locating each individual incident and showing increasing detail as you zoomed in on each point. It makes each incident personal, while making the dataset as a whole horribly compelling, inviting viewers to move across the map exploring each one.

What was the impact of your project? How did you measure it?

In November 2017 we launched Migration Trail, running it as a ten day event where everyone who saw the project was experiencing the same thing at the same time. (It is now available online in an 'on demand' version, where when a new viewer comes to the site it will start at day 1 and then run for the following ten days). Over the ten days between 20 November and 29 November (when the first ten day run of the experience ended), approximately 2,360,892 people saw the work, via the website itself, on social media (including partners) or in the press. 4,390 people visited the site, with 15% visiting multiple times. Visitors spent an average of 10.46 minutes on site. Using the Facebook Messenger bots that we built to deliver the instant messaging feeds we retained 67% of our audience over the ten days. On Facebook the @migrationtrail page reached 48,000 people directly, generated 5,289 post engagements (including 692 likes, 38 comments, and 114 shares) and received 5,418 video views over the ten days. On Twitter in the same period we had 65,270 impressions from the @migrationtrail account, with 1,996 engagements. We focused on measuring audience recruitment, retention and engagement over the ten days of the project and developed a REAN (reach, engage, activate, nurture) framework to set objectives for these categories and identify key performance indicators for each one. Duration was a key characteristic of this project and so retention of our audience was particularly important. Our channels for reaching our audience were: the main migration trail website; the Facebook bots for the instant messaging feeds; our email newsletter; Facebook and Twitter. We used Google Analytics and the inbuilt analytics tools of the platforms that we used to collect this audience data. We also collected qualitative feedback via user comments on Facebook and Twitter and a Google survey sent out after the ten days, which was very positive.

Source and methodology

There were two main methods for collecting the data needed: firstly, the team's field research and primary interviews; secondly, open source databases on migration, demographics, weather etc.Primary research: This was carried out over a two year period from August 2015 to July 2017 and involved a series of overland journeys, following the main routes taken by people on the move. In August 2015, I travelled from Lesbos, through Greece, the Balkans, Hungary, Austria, Germany, to the Netherlands. In June 2016 I spent two weeks on a rescue boat off the Libyan coast, in August 2017, I travelled to Turkey, Lesbos, Greece and Bulgaria, and in April 2017, to Sicily, through Italy to France. These journeys gave us first hand knowledge of many of the practical elements of these journeys which was essential for creating the detail of this project. Along the way, interviews were carried out with migrants, volunteers, NGO workers, local government, border police, academics and policy makers, to better understand the reasons people make these journeys, how they feel about it, what happens along the way and why these journeys play out as they do. We also followed a number of migrants, from Syria, Ghana and west Africa over a two year period, meeting them at several points along their journeys and staying in contact as they attempted to claim asylum and establish new lives. This material was also used to produce a podcast series to accompany the data visualisation.Secondary sources:The data in the visualisation came from a number of open source databases: European Commission (demographics, asylum claims, relocation quotas); United Nations High Commissioner for Refugees (numbers of refugees and asylum claims worldwide, countries of origin); Internation Organisation for Migration's Missing Migrants Project (migrant deaths, cause, date and precise location); International Labour Organisation (unemployment statistics); Passport Index (visa requirements); Journalism ++'s Migrant Files (border security spending); Open Sky Network (flights over Europe); ECMWF (wind); OpenAIS (shipping traffic).Fact checking: facts were corroborated using media reports, reports from organisations such as Amnesty International, the UN. We cross-checked the details of migrants' journeys using online travel planners on Google maps and Rome2Rio.

Technologies Used

The website was built using javascript, with the server integration done using PHP. We used the Leaflet.js library, with base maps from Open Street Map and map styling done in Mapbox.To create the animated routes and messages of the refugee journeys the process was to write a description of the route with rough timings (eg leaves Izmir at 20:30, arrives at coast 00:30) and a description of each place/part of the journey, to allow the writers to draft a narrative. At the same time, I drew a detailed route in Google earth. Pins showed stopping points (train stations, places people stop for the night), changes of transport or speed (the rubber boat starting to drift). This was exported to AutoCAD, paths tidied up and data extracted from (eg distances). Those distances, plus online research into timetables allowed us to work out how long the different parts should take. It was converted to GeoJSON, with stop/start times for each section of the route added to the code. This allowed the journey path to be animated. Route information was also transferred to excel, to create the main database. The excel sheets are broken down into ten second intervals - that's how often the server is called to update migrants' coordinates and data on eg phone battery. We built outwards from the migrants' location, to show when they would be likely to have access to wifi, or electricity to charge their phone. The writers' messages were also added to this database. We worked back and forth with the writers to refine the messages and ensure they accurately reflected where the character was and the conditions at any given time.Statistical maps were created using a database built by the team and leaflet. Additional drawn data (border walls) was drawn in Google Earth and AutoCAD. Detailed datasets with geolocation data (migrant deaths) was mapped using Mapbox. The animated datasets (wind, planes, boats) were geoJSON files integrated using leaflet. The Facebook bots were built using chatfuel.

Project members

Creative director: Alison KillingProducer: Josie Gardiner, Alison Killing, Michelle FeuerlichtFirst Prototype:  Mike RobbinsUX & Web design: Thomas LievestroTechnical realisation: Thomas LievestroAudio editor: Anik SeeAssistant producer: Sarah SaeyGraphic designer: Asja KeemanSound design and music: Bora YoonFixing and translation: Omar Shamil Mohammed, Nayief SalamehImpact producer: Michelle ChakkalackalData analyst: Luke GilderWriters: Elnathan John, Nadia Asfour, Zarghuna KargarPodcast narration: Marnie ChestertonAdditional fact-checking: Nayief Salameh, Benjamin Thomas White


Additional links

Project owner administration

Contributor username


Click Follow to keep up with the evolution of this project:
you will receive a notification anytime the project leader updates the project page.