NEVAC — Boat.viz

VAST 2008 Challenge
Mini Challenge 2:  Migrant Boats (geo-temporal analysis)

Authors and Affiliations:

Michael Stryker, The Pennsylvania State University, mzs114@psu.edu [PRIMARY contact]

Ian Turton, The Pennsylvania State University, ijt1@psu.edu

Chris Weaver, The Pennsylvania State University, cweaver@psu.edu

 

Student team: NO

Tool(s):

Improvise is a desktop application for building and browsing a wide range of flexible and powerful visual analysis tools. Live design of visual queries facilitates fast and flexible interactive drill-down into fine-grain relationships buried in spatiotemporal and social network information spread across multiple data sets. Cross-filtering queries across multiple views provides analysts with the means to seek out and dissect subtle patterns in complex information spaces.

 

We used Improvise to build boat.viz, an interactive visualization of the boat encounter data. The interface enables analysis of the pattern of encounters and the groups of people involved, using multidimensional drill-down into any combination of spatial, temporal, categorical, and nominal dimensions. (Click the preceding two links above to get the visualization and instructions for running it.)

 

 

Two Page Summary: NO

 

 

ANSWERS:

Boat-1 Characterize the choice of landing sites and their evolution over the three years.

Detailed Answer:

 

Click here to download the video that accompanies this section.

 

The reported landing sites for Isla del Sue_o migrants shows an observable progression in the number and diffusion of events over the three time period. In the first year, 2005, a small number of landings are concentrated along the Florida Keys with one event reported further north on the west coast of the peninsula near Fort Myers and few extending to the east to Fort Lauderdale. By 2006, landings are still focused on the Keys with the addition of events extending northwards along the western Florida coast north of Tampa. There are no events reported on the east coast north of Miami. By 2007, the Keys see proportionally fewer events and the west and east coast show a relatively even distribution of events to roughly the same latitude just north of Cape Canaveral on the Atlantic and north of Tampa on the gulf side. Notable gaps with no events across the three time period is a region south of Tampa and on southern tip of Florida where the Everglades National Park meets the coast. It is to be expected that the keys would create a shadow effect for landing events given that boats seeking this destination would have to first pass by or sail around the Keys. Another notable pattern in the landing destinations is the emergence of the Yucatan peninsula as the favored destination by year three. These sites account for roughly half of all landings.

:rsrc:Year1_landing.png

 

Figure 1: Landing sites in 2005

 

 

:rsrc:Year2_landing.png

 

Figure 2: Landing sites in 2006

 

 

:rsrc:Year3_landing.png

 

Figure 3: Landing sites in 2007

 

 

Within year variation:

Across all years the summer months we see a peak in landing events with the most activity occurring between April and September. This pattern is the most pronounced in the second and third year.

In 2005, landings occur almost exclusively on the Florida Keys. The first event as far north as Miami Beach appears in June. By 2006, the first event on the main land gulf-side occurs near Fort Myers in January. Landings still occur predominantly in the Keys though by June more events are observed near Fort Myers. It is in June 2006 that the first event is seen in Tampa and in July the first event north of Tampa. By October more are occurring south of Ft. Myers along the western coast of Florida in the Naples area though a few occur in the Everglades National Park. The first events near Cancun, Mexico occur in July and September events in 2006. This becomes a very popular destination by 2007, with at least one voyage per month from February to September. Though entirely absent in the first year, Mexico accounts for roughly half of all reported landings in the data set. In February of 2007 the first event near Cape Canaveral/Canaveral National Seashore is reported. By April, events can be seen just as far north on the west coast near Homsassa and Port Richey to the South in what appear to be less developed areas.

 

Visualization techniques:

(1) Animate landing sequence defining a temporal filter of several months and sliding this filter over the time line. The effect is to animate the sequence of landing events on the map.

(2) Set the temporal filter at an interval of one year. Observe annual pattern.

(3) Using the reruns calendar view, set cycle to seven days and look for pattern by day of the week.

(4) Reruns view to look for seasonality. Summer peak in mid-July in all three years.

(5) Weekday pattern: Sunday only slightly favored. Even distribution throughout week of total events. Landings slightly more prevalent on Sunday, Thursday, and Friday.

 

By vehicle type:

The pattern of rustic vehicle type landings dominates that of the overall data set. This type is favored by the migrants four to one over the _go fast_ and three to one over _raft_. As expected by their names, the raft events generally display a shorter range, though they still reach land in Cancun, near Tampa and near Cape Canaveral. The _go fast_ vessel types land in all of the same areas as the other two vessel types but are disproportionately land more frequently in the northern-most extent of the west and east coast of the Florida peninsula. By year three the _go fast_ vessels appear to be avoiding the keys and focus landings near Tampa and Cape Canaveral on the upper east and west coasts. The rustic shows a similar pattern though less distinct favoring the main land on both coasts over the keys. The raft still shows a more even distribution of events across the peninsula, the keys, and the Yucatan peninsula. Still, by the second half of 2007, only one raft landing is reported for the keys.

 

Boat-2  Characterize the geographical patterns of interdiction over the three years

Short Answer:

The interdiction patterns reflect the patterns evident from the landing choices over the three years. In year one, interdiction events occur almost exclusively in a north-south corridor between the Isla del Sue_o and Florida roughly the width of the island. Despite numerous landings in the western Florida keys, very few interdictions occur south of the western half of the keys. By year two, interdictions occur densely clustered off the east coast of Isla del Sue_o and now are spread evenly along the length of the keys. A pocket of interdictions now occurs in the region bound by a line from Ft. Myers to the western most key and the mainland coast to the east. A few also appear on the Atlantic coast near Palm Beach. By year three, interdiction coverage has spread in a band arcing from Palm Beach past to the western most keys. A set of events now fully surrounds the island. The alcove between Ft. Myers and the keys no longer shows interdiction events. Instead a line of events extends roughly northward from western key past Tampa Bay. Interestingly, no interdictions occur on the Atlantic side north of Ft. Lauderdale.

 

:rsrc:Year1_interdiction.png:rsrc:Year2_interdiction.png:rsrc:Year3_interdiction.png

 

Figure 4.  Interdiction geographic pattern for 2005-7 (from left to right)

 

 

Boat- 3 What is the successful landing rate over the time period?

Short Answer:

The landing rate for the data available is nearly fifty percent by vessel and surprisingly consistent across each type. There were 441 landings (66 Go Fast, 88 raft, 287 rustic) out of 917 documented voyages for a success rate of 48%. Not all of these landings benefitted from the US _wet-foot-dry-foot_ policy. One hundred ninety landings occurred in Mexico, reducing the success rate overall to 27% if the goal had been to reach US soil on all voyages. That may have been true for some vessels that eventually landed in Mexico, so the successful rate of seeking asylum is somewhere between 27% and 48%. For those records that included a launch location, the combination of launch and apparent landing destination revealed some interesting patterns for success and failure.

By launch location:

There are five general launch locations on the island. Two are in the northwest corner, a third just east of the northern most, the fourth and fifth on the two prominences on the eastern coast (the jut outwards towards the Atlantic). Using the technique of filtering by year and then mousing over an inset map of the island with a radial point buffer, a direct path for the highlighted events is displayed on the main map. In year one, the southern-most launch site shows almost all voyages were interdicted and the western most site shows any voyage eastward was interdicted. The center launch location shows the most success. Year two shows the same distinct failure rate for the east-bound boats from the western most launch point. The middle site is still the most successful but somewhat less so.