Analyzing Vessel Movements
Based on the data we acquired for vessel movements in the mediterranean, I’ve created a series of maps that target the vessels off the coast of northern Libya, especially near Tripoli and Zuwara, where the Ghostboat most likely departed from. They’re grouped by category which is represented by the “AIS type” in the original data.
The data is way too big for a google spreadsheet so I ended up importing it into an SQL database. This made it a lot faster and easier to query specific information. But my next thought was, where do I even start?
Clues from the “Left To Die Boat”
A couple weeks ago, Ross Whiteford mentioned a story similar to the Ghostboat about a group of refugees who set out from northern Libya, bound for Italy, where only 9 (out of 72) survived despite authorities in the area knowing where they were. It came to be called the “Left-To-Die boat” and was extensively researched by Forensic Architecture who published an entire case study on it. You can read it in full here: The Left-To-Die Boat, Forensic Architecture. This case was extremely helpful in figuring out where to look in the massive table of data.
In the above image, taken from the Left-To-Die Boat case study, the red line traces the refugees path as they head out to sea. After about 30 hours of travel they run out of fuel, making it approximately 180-200km off the coast of northern Libya at which point they are adrift. This gave me an idea of how far a boat could travel in a given amount of time, therefore giving me an idea of what kind of geographical area to search for vessels. Obviously, variables still exist. I don’t know what kind of boat the Ghostboat was or if they took a similar route. But it was something to go on.
The Search Area
Pictured above, I decided to narrow my search to vessels south of 35º north and between 11.5º and 14º east. The Ghost boat would have had to travel through this area when it left the Libyan coast.
Update: I just learned, according to a witness statement, that the Ghostboat left from the “Tripoli area”. Before, we weren’t sure where exactly it left from so the area around Zuwara was included.
The Maps
The following maps trace all the vessels that passed through the search area between midnight on June 27 and midnight on June 29 (48 hours). About 50 in total. They’re separated according to vessel type, which is indicated by the “ais” code in the data. Each colored line represents a unique vessel. Each point on a line is the location of that vessel at a point in time. By clicking on a point you can see more information about the vessels status as well as a link to the VesselFinder.com page for a full profile.
Note: For the most part, I decided not to include vessels that stayed in port during these 48 hours. My guess was that the Ghostboat would have tried to avoid ports or busy areas as it set out on its journey.
2 Italian Coast Guard Boats (AIS: 55)
These 2 vessels were probably the most interesting in their movements. They seem to have covered the same territory at different times. If the Ghostboat made a distress call, these are likey the guys who would have come to help. In the bottom left area of their movements CP906 stays put in an area for 12 hours, then leaves. Datillo CP940 revisits the exact same location 24 hours later. Perhaps they were investigating something?
http://bit.ly/1Nf9rkI
3 Fishing Boats (AIS: 30)
11 Tug Boats (AIS: 52)
15 Cargo Vessels (AIS: 70–79)
13 Tankers (AIS: 80–89)
6 Vessels classified as “Other” (AIS: 90–99)
Moving Forward
My hope is that these maps provide some leads into vessels worth contacting about their operations and/or encounters at sea between June 28 and 29, 2014.
If you’re interested in investigating the data for yourself, I would suggest setting up a database on your computer and importing the data into it. Web-based tools like google spreadsheets can’t handle it all. I used SequelPro (for Mac) to create an sql database that I was able to query with sql statements. Workbench seems to be a similar tool available for Windows.
For mapping, I’ve been using CartoDB which has software enabling users to import geographical data into a table and style it with on a customizable map.
