Sandia National Laboratories / Kitware - DatabaseView

VAST 2007 Contest Submission

Authors and Affiliations:

Patricia Crossno, Sandia National Laboratories, pjcross@sandia.gov

Brian Wylie, Sandia National Laboratories, bnwylie@sandia.gov

Andrew Wilson, Sandia National Laboratories, atwilso@sandia.gov

John Greenfield, Sandia National Laboratories, jagreen@sandia.gov

Eric Stanton, Sandia National Laboratories, etstant@sandia.gov

Tim Shead, Sandia National Laboratories, tshead@sandia.gov

Lisa Ice, Sandia National Laboratories, lgice@sandia.gov

Ken Moreland, Sandia National Laboratories, kmorel@sandia.gov

Jeff Baumes, Kitware, jeff.baumes@kitware.com

Berk Geveci, Kitware, berk.geveci@kitware.com

Student team: [ ] YES [ x ] NO
If you answered yes, name the faculty who agreed to be your sponsor: Name, email address

Tool(s):

We used "Clear Forest" (http://www.clearforest.com) - a commercial product - for doing entity extraction and building relationship tables. The VAST 2007 raw data set was imported into the Clear Forest tagging engine and the resulting entities and relationships were exported as XML. The contents of the XML were then imported into a MySQL database using a custom parser and a schema of our own design.

The TITAN toolkit (http://www.vtk.org), developed by Sandia National Laboratories in collaboration with Kitware, was used to create a custom tool - "Database View" - that read the entity / relationship data from the database for display using a variety of graphing techniques. Users interact with DatabaseView by creating SQL queries for nodes, edges, and selections - which are displayed as interactive graphs with control over layout, color, labeling, etc. Interactive selection can be used to display subsets of the main graph in a separate view with its own layout and display parameters. Although it required a degree of SQL expertise, building graphs in this manner through arbitrary SQL queries gave us the flexibility and control needed for rapid exploration of complex queries.

Provide a short description of the tool(s) you used. Mention where and when it was developed.
Additional credit to developers of the tools can be provided here, and links to find more information on the tool.

(250 words MAX)

Data set used: [ x ] RAW DATA SET [ ] PRE-PROCESSED SET

TOC: Who – What – Where – Debriefing - Process - Video

---

1. WHO: who are the players engaging in questionable activities in the plot(s)? When appropriate, specify the association they are associated with

Name	Associated organization	Involved in illegal activities? (Yes/No)	Involved in terrorist activities? (Yes/No)	Most relevant source files (5 MAX)
Cesar Gil	Gil Breeders	Yes	Yes	Chinchilla Dreamin', Week-of-Mon-20040705.txt_86, Week-of-Mon-20030901-1.txt_36, Week-of-Mon-20030609.txt_4
Faron Gardner	Animal Justice League (AJL)	Yes	Yes	Week-of-Mon-20030602-1.txt_66, Week-of-Mon-20030609.txt_4, Week-of-Mon-20030818.txt_23, Chinchilla Dreamin'
Catherine Carnes (nickname Collie)	Society for the Prevention of Mistreatment to Animals (SPOMA)	Yes	Maybe	Week-of-Mon-20030526-2.txt_57, Week-of-Mon-20031013.txt_4, Week-of-Mon-20030818.txt_23, Chinchilla Dreamin'
Luella Vedric	SPOMA	Maybe	Maybe	Week-of-Mon-20030526-2.txt_57, Week-of-Mon-20040412-2.txt_13, Week-of-Mon-20031013.txt_4, Week-of-Mon-20040119-1.txt_98
Rapper r’Bear (aka r’Bert – assuming typo in news)	SPOMA; Shraavana (aka Shravaana)	Yes	No	Week-of-Mon-20030609.txt_7, Week-of-Mon-20040119-1.txt_98, Week-of-Mon-20040308.txt_109, Week-of-Mon-20040614.txt_94, Week-of-Mon-20040628.txt_61
Madhi Kim	Global Ways	Yes	No	Week-of-Mon-20040412-2.txt_13, Week-of-Mon-20040308.txt_109, Week-of-Mon-20030526-2.txt_57
Abu Hassan (aka Professor Assan)	Global Ways; Assan Circus (aka Professor Assan and his Amazing Animals)	Yes	No	ImportPermitsv3 BEST WORKING COPY, Week-of-Mon-20031215-1.txt_91, Week-of-Mon-20040301-1.txt_75, Week-of-Mon-20031013.txt_4
MN		Yes	No	meeting
Rosalind Baptista		Yes	No	hunt8, meeting, 20040630

(Note: ignore the file extensions of the files)

---

2. WHEN /WHAT: What events occurred during this time frame that are most relevant to the plot(s)?

Provide a text list of events following the sample layout. Use short description (i.e. one or 2 lines per event)

Provide what you think is the best subset of events (20 events MAX)

	Date Can be a range	Event description	Most relevance source files (5 Max)
1	2001-2002	Mary Allen Ollesen (Westminster, CA) got several Gambian rats when their owners gave them up. “They get sick and die, and you don’t know why.”	Week-of-Mon-20040705.txt_86
2	3/1/2003 – 3/1/2004	Import permits obtained where Abu Hassan acts as consignee for Global Ways to import animals from various African countries into the US (New York).	ImportPermitsv3 BEST WORKING COPY
3	5/31/2003	Luella Vedric at a SPOMA fundraiser says she is good friends with SPOMA director Catherine Carnes, dislikes violent animal rights groups, and gets her tropical fish from Mr. Kim.	Week-of-Mon-20030526-2.txt_57
4	6/5/2003	Three PetSmart stores in Los Angeles vandalized and back room animals abducted by Animal Justice League (AJL). Faron Gardner wanted. Cesar Gil’s response to raid defends Gardner’s actions.	Week-of-Mon-20030602-1.txt_66, Week-of-Mon-20030609.txt_4
5	7/16/2003	AJL sends beef and letter to Los Angeles Times, claiming meat poisoned in 20 LA supermarkets. No poisoned meat found.	Week-of-Mon-20030714-2.txt_25
6	7/18/2003	Rosalind Baptista photographed poaching chinchillas in Choapa Valley, Chile.	hunt8
7	8/15/2003 - 9/1/2003	Cesar Gil starts Gil Breeders, a chinchilla farm. Sells chinchillas at the West LA Farmer’s Market every Sunday, 9am-2pm.	Chinchilla Dreamin', Week-of-Mon-20030901-1.txt_36
8	10/2003-1/2004	Global Ways’ tropical fish shipments from South America have contaminated packaging that makes people sick. Symptoms include numb/tingling hands, dialated eyes, breathing difficulty, and euphoria (similar to cocaine).	Week-of-Mon-20031027.txt_57, Week-of-Mon-20040105-1.txt_58, cocaine hydro, Transport of Live Fish, DEA Files Updatev2
9	1/17/2004	SPOMA dinner at Millennium Broadway Hotel hosted by Luella Vedric nets $230,000. Rapper r’Bear performs and donates $80,000.	Week-of-Mon-20040119-1.txt_98
10	10/13/2003 – 3/2/2004	Hassan uses Assan Circus as traveling cover while trapping and illegally smuggling protected wildlife from Africa to US. Collie Carnes of SPOMA is working with CITES and Luella Vedric to track him.	Week-of-Mon-20031013.txt_4, Week-of-Mon-20031215-1.txt_91
11	3/2/2004	Animal Defenders International secures CITES confiscation order to seize Assan Circus animals in Zimbabwe. Hassan flees the country.	Week-of-Mon-20040301-1.txt_75
12	3/14/2004	Madhi Kim tours Shravaana Exotic Animal Sanctuary and spends day with rapper r’Bear.	Week-of-Mon-20040308.txt_109
13	4/2004	Rosalind Baptista meets with MN along Bourbon Street in the French Quarter in New Orleans (we are assuming that RB is Baptista since the photo was grouped with the poaching images).	meeting
14	4/2/2004	Cesar Gil posts cartoon to blog showing a chinchilla being infected by exposure to a small rat-like animal.	Chinchilla Dreamin', 20040402
15	4/17/2004-4/18/2004	Global Ways hosts auction "Nights of Champagne and Tropical Fish" at the Miami Beach Convention Center. Luella Vedric and r’Bert are special guests of Global Ways owner Madhi Kim on Saturday night.	Week-of-Mon-20040412-2.txt_13
16	4/2004 – 6/20/2004	Rapper r’Bear adds 500 new animals to Shravaana, including short-tailed chinchillas.	Week-of-Mon-20040614.txt_94
17	6/30/2004	Cesar Gil posts cartoon to blog saying that Senorita Baptista will be delivering infected poached chinchillas to US customers.	Chinchilla Dreamin', 20040630
18	6/30/2004	r’Bear taken to hospital with bumpy face and fever. Shravaana Open House cancelled.	Week-of-Mon-20040628.txt_61
19	7/7/2004	Monkeypox outbreak hits Los Angeles area with seven people seriously ill, 50 possible cases. Chinchillas may be carriers of a strain similar to that found in Africa (1% to 10% death rate).	Chinchilla Dreamin', Week-of-Mon-20040705.txt_83
20 max	7/24/2004	Two dead in monkeypox outbreak. Cesar Gil sought and believed to have fled country.	Week-of-Mon-20040705.txt_86

---

3. WHERE: What locations are most relevant to the plot(s)?

Follow this example layout. Use only one-line per item.

	Location	Description	Most relevance source files (5 Max)
1	Los Angeles	Site of monkeypox outbreak and AJL terrorist incidents.	Chinchilla Dreamin', Week-of-Mon-20040705.txt_83, Week-of-Mon-20040705.txt_86, Week-of-Mon-20030602-1.txt_66, Week-of-Mon-20030714-2.txt_25
2	Shravaana Exotic Animal Sanctuary (near San Diego)	Madhi Kim meets with r’Bear, many endangered species added to animal sanctuary, and r’Bear cancels open house when he gets sick.	Week-of-Mon-20040308.txt_109, Week-of-Mon-20040614.txt_94, Week-of-Mon-20040628.txt_61
3	New Orleans	Baptista meeting with MN	meeting
4	Miami	Contaminated packaging in tropical fish shipments, Global Ways fish importing offices, and “Nights of Champagne and Tropical Fish” auction.	Week-of-Mon-20031027.txt_57, Week-of-Mon-20040105-1.txt_58, Tropical Fish Importers, Week-of-Mon-20040412-2.txt_13
5 max	Choapa Valley, Chile	Chinchilla trapping location and probable initial chinchilla exposed to monkeypox.	hunt8

---

4. DEBRIEFING

Include your written assessment of the situation (between 1000 and 2000 words)

This narrative should describe the plot(s) and subplots(s) and how people, motivations, activities and locations are part of the plot. Include in your narrative the relationships of the various players. If there are uncertainties, you can suggest possible next steps to clarify those uncertainties.

(NOTE: here there is no need to explain how the tool helped you, focus on convincing us that you UNDERSTAND the situation).

Main Plot

The main plot is to create a monkeypox outbreak carried by pet chinchillas to halt demand for, and poaching of, endangered wild chinchillas. The central actor in this plot is Cesar Gil, a chinchilla lover, biologist, animal rights activist, and blogger. The plot and its motivation are described in a series of cartoons called ‘Chinsurrection’ on Gil’s blog, Chinchilla Dreamin’. News articles, with dates corresponding to those in the blog, confirm that the plot is executed in the Los Angeles area leading to two deaths and numerous illnesses. Gil is sought in connection with the outbreak. Gil has a chinchilla farm in the Los Angeles area called ‘Gil Breeders’ and sells chinchillas at the West LA Farmer’s Market on Sundays. The ad listing chinchilla types includes short and long-tailed varieties, both of which are endangered. Since domestically farmed breeds are a cross between long and short-tailed chinchillas and only animals born in captivity from parents born in captivity can be sold legally, these chinchillas must have been obtained from a poacher. Although Gil is vehemently against poaching, his website makes clear that he is willing to sacrifice a few animals to save the rest.

The monkeypox strain seen in Los Angeles was different than that seen in the outbreak in the mid-west. No one died in the mid-west, while two people died in Los Angeles. So prairie dogs or animals from the mid-west were not the source. The lethality of monkeypox in African outbreaks suggests an African source; the most recent outbreak was in the Congo. None of the animals coming into the USA legitimately through the import permits came from the Congo. Plus, the only importer into the USA was Global Ways, which was the sort of smuggling operation that Gil sought to shut down. So they would not have participated. Although there were news items about animal research lab break-ins, they were outside of the Los Angeles area and no connections could be made to monkeypox. Nor could a connection be made between Gil and the perpetrators or their organizations. Taking a clue from Chinsurrection, the animal infecting the first chinchilla with a sneeze looks like a rat. Mary Ann Ollesen’s statement that she had gotten several Gambian rats a few years ago when their owners gave them up, and that they gotten sick and died, suggests that Gambian rats could have been the source. Ollesen has her business in Westminster, which is in the Los Angeles area. As a biologist living in the Los Angeles area, Gil could have come into contact with an infected rat, known how to handle the virus, and hatched his plan. Questioning Ollesen about whom she received the Gambian rats from, and then questioning those people in turn, could provide leads into Gil’s source.

There are varying degrees of certainty about the players in various plots. Although Gil’s role is certain, the roles of his friends Collie and Faron are less clear. Through news articles, we find that they are ‘Collie’ Catherine Carnes, director of the Society for the Prevention of Mistreatment to Animals (SPOMA), and Faron Gardner, spokesperson for the Animal Justice League (AJL). Although Gardner does not appear to be part of the monkeypox plot, he is involved in the Animal Justice League subplot. Gardner is quoted as having overheard Collie say that getting arrested is fun, so Carnes and Gardner also know one another, and all three players are willing to break laws to further the animal rights movement. Carnes may be involved, because she is the first step in a chain of relationships which could be used to connect Gil with a chinchilla poacher. However, Carnes may not know the ultimate goal of the plot, since monkeypox is not identified on Gil’s blog until the outbreak has already happened. She may believe that she is involved in a sting operation to stop poaching and wild animal smuggling. Both Gardner and Gil live in the Los Angeles area, but it is unclear where Carnes lives.

Continuing to explore the web of relationships expanding from Gil, Collie Carnes is good friends with Luella Vedric, a New York socialite who is a major contributor to SPOMA, and who hosted a SPOMA fundraising dinner at the Millennium Broadway Hotel. Vedric collects tropical fish and her supplier is Madhi Kim, owner of Global Ways, a company which imports both tropical fish and animals from overseas to cover their drug smuggling and endangered species smuggling operations (see subplots below). The tropical fish are imported from South America and the animals are from Africa. Vedric and Kim both have contact with the rapper r’Bear, who is involved with a wild animal sanctuary for endangered species, Shravaana (or Shraavana). Kim visits Shravaana in March 2004. Between April and June, 500 new animals are added to the sanctuary, suggesting that Kim’s visit was to setup a deal for Global Ways supply those animals, including short-tailed chinchillas. In Chile, Rosalind Baptista is a chinchilla poacher. Assuming that R.B. stands for Rosalind Baptista, she meets with MN in New Orleans in April 2004, presumably to arrange smuggling of poached chinchillas into the US. Although we have been unable to establish who (or what organization) the initials MN stand for, the timing is right for Baptista to be the source of r’Bear’s chinchillas. Gil names “Senorita Baptista” as a source of smuggled chinchillas in Chinsurrection. Assuming that the details in the cartoon are accurate, Gil knows how many chinchillas Baptista is smuggling and when they will arrive, showing that he has a source of inside information on her activities. These people fall into two adversarial groups: the wildlife smugglers / poachers and the animal rights activists. Luella Vedric is on the boundary between the groups, but her allegiance is clearly with the animal rights crowd. Despite Vedric’s seemingly friendly association with Mr. Kim, her work with CITES and Collie Carnes / SPOMA to track down and stop illegal wildlife trafficking by outfits like the Assan Circus indicates that her motive for getting close to Kim is probably to get information as part of a sting operation (see animal smuggling subplot below).

Although the rapper r’Bear appears similar to Vedric in his relationship with Kim, r’Bear falls into the smuggling group because he is illegally purchasing endangered wildlife. This is established by r’Bear’s new short-tailed chinchillas at Shraavana, followed by the open house cancellation when r’Bear is hospitalized with symptoms resembling monkeypox. Since r’Bear’s chinchillas are short-tailed and no short-tailed chinchillas have entered the country through any of the import permits, these chinchillas must have been illegally obtained. However, this presents a problem in explaining how r’Bear’s chinchillas get infected, since he is near San Diego and he has a direct connection to Global Ways, but no direct link to Cesar Gil. In fact, there are really two different outbreaks. r’Bear’s hospitalization happens a full week before the outbreak in Los Angeles is reported. So although the farmer’s market could have provided a means for distributing infected chinchillas to others in the Los Angeles area, the infection of r’Bear’s chinchillas requires access (or someone with access) to the chinchillas during the poaching / smuggling portion of the delivery pipeline.

The person with the best access would be the photographer who took pictures of Rosalind Baptista in the Choapa Valley, Chile. Knowing where the traps are located would allow the photographer to infect one of the trapped chinchillas before Baptista collected it. That one chinchilla would then infect all the chinchillas in that shipment, which closely resembles the scenario shown in Chinsurrection. Although we have no evidence to connect the photographs with any of the players, logical candidates would be either Carnes or Vedric since they are both involved in investigating and stopping wild animal poaching and smuggling. Vedric can probably be eliminated since she has stated that she’s “against violence toward any animal and people are just another kind of animal”. So, unless Vedric is lying, she does not seem likely to infect either a chinchilla or a person with monkeypox. Also, Gil’s detailed and timely information about Baptista is best explained by Carnes being the source, since they have a direct link. On November 7, 2003 Gil writes, “…and now, the trappers are back in South America. Don’t ask how I know, I just do.” Since the photographs of Baptista were dated in July 2003, they could have been the source of his statement.

Animal Smuggling Subplot

Abu Hassan is using the Assan Circus as a cover for illegal wild animal trapping and smuggling as the circus travels throughout Africa. Hassan acts as the consignee for Global Ways to import animals from all over Africa (Kenya, Uganda, Zambia, Tanzania, and Zimbabwe) into the United States in a series of import permits issued between March of 2003 and March of 2004. Endangered species are hidden among permitted animals or dangerous species so that inspectors will be less likely to hand-inspect shipments. All of Hassan’s permits included lions, and all but one included tigers. In October of 2003, Carnes and Vedric were working on tracking and stopping the operation. In March of 2004, Animal Defenders International gets a CITES confiscation order from the Ministry of Agriculture on animal welfare grounds and shuts down the circus when it is in Zimbabwe. Hassan flees the country.

Drug Smuggling Subplot

Global Ways is using shipments of tropical fish to smuggle drugs into the US from South America. The Fish and Wildlife Service (FWS) issues a warning not to handle shipments of tropical catfish from South America that entered the US through Miami. Three Florida fish retailers have had sick employees after handling the packaging. Although Global Ways is only one of a list of businesses being investigated as the source for the shipments in the FWS warning, Global Ways is clearly identified as the single source in a number of letters to the editor. These complain about low fish survival rates and shipping bags covered in some noxious substance that causes tingling of the hands, dialated eyes, breathing difficulty, and possible euphoria. These symptoms match those listed in the material safety data sheet for exposure to cocaine hydrochloride, demonstrating that drugs are being smuggled in the packaging of the fish shipments. The combination of Figure 6 (Transport of Live Fish) showing foam rubber insulation with the Intelligence Alert on heroin impregnated foam backing from South America (DEA Files Update) suggests that foam rubber was used in a similar manner to smuggle cocaine.

Animal Justice League Subplot

The terrorist activities of the AJL in the Los Angeles area form a final subplot. Gardner is wanted in connection with three PetSmart break-ins, in which stores were vandalized and backroom animals were abducted. The AJL website states that this would not be the last attack and Cesar Gil is quoted as defending Gardner and the AJL’s actions. In another incident, the AJL leaves a package of beef and a letter at the offices of the Los Angeles Times. The letter claims that meat has been poisoned at 20 Los Angeles supermarkets, though no poisoned meat is ever found. Although we have not found anything directly linking Gil to these incidents, his relationship with Gardner and defense of the raids make it a possibility. Also, one news article states that in addition to being sought for the monkeypox outbreak, Gil is “a suspect in radical animal rights activities”.

---

5. VISUALS and Description of ANALYTICAL PROCESS

Data Preparation

We extracted tagged entities and relationship tables from the data by running it through Clear Forest, a commercial entity and relationship extraction tool, which outputs a series of XML files. These were then imported into an SQL database using a custom schema and parser of our own design. Additional links between entities were created by linking entities that appeared within the same document. Text within images was manually entered and entities were hand-tagged. Since these did not go through Clear Forest, there are some irregularities related to the integration of these images.

Our strategy is to use all entities, terms, source documents, and known relationships from the database to generate a graph that casts as wide a net as possible, and then to probe that graph with focused queries to generate subgraphs containing relevant pieces of the puzzle. We used a custom Sandia-developed tool, DataBaseView (see tool description above), to do this. The tool’s full interface is shown in Figure 1 below. Although the full interface has a large number of windows, unneeded ones can be hidden, while the others are expanded to make full use of the available screen real estate. A detailed explanation of the tool startup can be found in the video. Once we have connected to the database and generated our wide-net graph (a corner of which can be seen in the graph to the left of the one enclosed by the pink box), we do a focused query that returns a list of entities and terms (the list directly above the pink oval). Each of these is then individually selectable for inclusion (or exclusion) as relevant nodes in a subgraph. The subgraph is generated by taking a neighborhood of nodes and edges around each of the selected nodes from within the wide-net graph. A neighborhood of one includes all nodes directly connected to the selected nodes by an edge. A neighborhood of two includes all nodes that can be reached by traversing no more than two edges. Larger neighborhoods are not useful to generate because the volume of extraneous nodes and overlapping edges hide whatever useful information may be represented in the graph. The neighborhood is set through the parameter field in the pink oval. The button to its left activates the graph generation. Alternatively, nodes can be selected through a non-SQL interface (green box) that lists Clear Forest tag categories in one window, and the values for the selected category within the other. Within the generated subgraph, source files can be selected (both text and image) through rubber-banding. The selected files are then listed in the window shown in the upper right corner of the interface. An image (blue box) is displayed below (bottom of blue arrow). Although a great deal can be seen in just the entity/term relationships shown in the subgraph, much of our analysis involved reading articles or viewing images whose relevance was suggested by the subgraph.

Figure 1: DataBaseView showing the full interface with all query windows, neighborhood distance (pink ellipse), image source file selection (blue box) with associated image display (blue arrow), and the solution graph (pink box).

Our general approach was an iterative one. Starting from some conceptual thread or idea, we perform the following steps repeatedly to extract an expanding set of entities and relationships to form a solution graph:

· identify a thread to investigate

· do an SQL query on entities and terms involved in that thread

· select relevant entities and terms from a list returned by the query

· select a neighborhood distance value

· generate a graph containing the query result

· select source documents in the graph for examination

· view source documents to identify next query thread

In the following discussion, we present a mostly linear path that attempts to cover our plot discovery process. However, in reality, our discovery process was highly non-linear. We explored many dead-ended theories that we will spare you from hearing about. We revised our ideas about the relationships between people and their roles. However, we did immediately latch onto the idea of a monkeypox plot because we came upon the term while we were manually entering the text from the Chinsurrection cartoons. That led us to our first query thread, which was monkeypox. By this, we mean that we constructed an SQL query asking for all entities or terms in the database containing ‘monkeypox’. The resulting graph using a neighborhood of one was disjoint and not very interesting, so we expanded to a neighborhood of two, which generated the graph shown in Figure 2.

Figure 2: Closeup of solution graph for monkeypox query with a neighborhood of two. Notice that entity nodes are shown in red, text-based files are green, images are cyan, and terms are in various shades of blue.

In Figure 2, the original Chinsurrection image (20040707.jpg) appears in the lower part of the graph connected to the term monkeypox (both combined in the bottom pink circle) and the Chinchilla Dreamin’ htm source file (circled in green). In the pink ellipses to either side of the green circle, we see the entities Los Angeles and Cesar Gil. On the far right and left edges of the image are two dense balls of entities and terms surrounding two news articles (Week-of-Mon-20040705.txt_83 and Week-of-Mon-20040705.txt_86). Within these dense balls are nodes for ‘Monkeypox’, ‘Pet chinchilla’, and ‘monkeypox outbreak’. Reading the two articles we find that the strain of monkeypox in the Los Angeles outbreak is a potent strain that seems similar to one found in Africa (Democratic Republic of the Congo). However, even without reading the articles, we can see from the graph that the monkeypox outbreak is somehow related to Cesar Gil, chinchillas, and Los Angeles because of the edges connecting those nodes to one another through various source files. Although additional articles pertaining to monkeypox were part of the graph (note the edge exiting the image on the right), they were not of interest.

Next we explore chinchillas as a thread and once again look at a neighborhood of two around the query results to get a broader view of things. The resulting graph is shown in Figure 3.

Figure 3: Query on chinchillas with a neighborhood of two.

The monkeypox outbreak is shown in the upper left portion of the graph (large pink circle). We also see references to poaching and Chile and Rosalind Baptista (pink circles). One thing to point out in Figure 3 is that we have connected meeting.jpg (green circle) to the Chinchilla Dreamin’ blog. Due to the manual text entry, the Rosalind Baptista and RB tags (which we assumed was Rosalind Baptista since the images were in the same directory) linked to images that were already connected to the blog. So we linked all of the images in that directory to the blog so they could be displayed, even though some of them lack text tags. At this point in the analysis, we can make new inferences and resolve duplicated entities. We have an entity resolution capability, whereby we can map various names to a single main entity description. The mapping between Rosalind Baptista and RB reflects this entity resolution step. The same is true of linking the meeting image to New Orleans. Thinking that this image looked like New Orleans because of the wrought iron balconies, we were able to find almost the same image through Google (http://www.jankaulins.com/aacgproduct.php?p=205), which enabled us to establish a definite location for the meeting.

Thinking that Rosalind Baptista might be part of the plot, we queried her, but nothing new was found. We also queried New Orleans, but everything we found outside of the meeting image seemed totally unrelated. We searched extensively for any entity (person, organization, or company) that had the initials MN, but every match seemed unrelated. So, we backed up to the thread suggested in Figure 2 and added Cesar Gil to the monkeypox query to get the result shown in Figure 4.

Figure 4: Cesar Gil added to the monkeypox query with a neighborhood of one.

We select all four of the source text files shown in the graph. These are listed in the upper right corner. Of these, we display the article Week-of-Mon-20030609.txt_4 in the window in the lower right corner. We have indicated the article’s location in the graph with the pink circle. In the graph, the article connects Cesar Gil and Los Angeles, but the text of the article is not about the monkeypox plot. Instead, we see the first hint of the AJL and the other animal rights terrorist activities in Los Angeles. Important terms (PetSmart, Faron Gardner, and AJL) are circled in green. Gil’s statement appears to be defending Gardner’s actions, so we decide to follow Faron Gardner as a possible co-conspirator. Although we follow this thread, note the number of times the term poaching appears in the lower left portion of the graph and how Chile (the red node above the poaching nodes) seems connected. So without looking at the source files (which are the Chinsurrection images) we are already making a connection between Chile, poaching, Cesar Gil and monkeypox through the Chinchilla Dreamin’ blog (the green node at the center of it all).

So now, we query on Faron Gardner and get the graph shown in Figure 5.

Figure 5: Solution graph after adding Faron Gardner to the query.

The new solution graph now includes the Animal Justice League (upper pink circle) within its neighborhood. Animal Justice League is linked directly to Faron Gardner (lower pink circle). We have selected three of the articles in the graph (the green nodes highlighted in white within the yellow circle) that are linked to either Faron Gardner or Animal Justice League. From the list of these selected articles shown in the upper right window, we have displayed Week-of-Mon-20030602-1.txt_66 for reading. Here we find the details of the PetSmart (middle green circle) raids and are able to connect AJL with the Animal Justice League (top and bottom green circles), which we use to do an entity resolution between the tags.

Seeking to expand this thread and look for connections between the AJL and the monkeypox plot, we add Animal Justice League and PetSmart to our query. However, a neighborhood of one does not give us any particular insights, so we expand to a neighborhood of two, a zoomed-in section of which is shown in Figure 6.

Figure 6: Solution graph showing addition of Animal Justice League and PetSmart.

We focus on the area (pink circle) near the Animal Justice League (blue circle), where we find the article Week-of-Mon-20030526-2.txt_57 connecting Catherine Carnes to AJL and the Society for the Prevention of Mistreatment of Animals. Selecting this article for reading, we discover that Catherine Carnes is nicknamed Collie (which connects her to Cesar Gil since he lists Collie as a friend in his blog) and that she is director of SPOMA. Additionally, the article further expands the social network of the animal rights activists by establishing a connection between Carnes and Luella Vedric, who says that Collie is one of her dearest friends. We link Collie and Catherine Carnes, along with SPOMA and the Society for the Prevention of Mistreatment to Animals, to our entity resolution table. Since Vedric is not within our neighborhood, we need to add her directly to the query. Plus, adding Catherine Carnes and SPOMA might further expand our social network or uncover more animal rights terrorist activities, so we add all three to our query. A closeup of the query result is shown in Figure 7.

Figure 7: Closeup of query results for adding Luella Vedric, Catherine Carnes, and SPOMA

By adding Vedric and Carnes to the graph (and reducing the clutter by reducing the neighborhood back to one), we can now see a fairly clear representation of the people (Gil, Gardner, Carnes, and Vedric), organizations (AJL and SPOMA), locations (Los Angeles and New York), source articles (Chinchilla Dreamin’ and a number of news articles), and the relationships between them in Figure 7. Selecting the two articles that link Carnes and Vedric (pink circle), we read Week-of-Mon-20041013.txt_4 and see that Vedric and Carnes are tracking the Assan Circus in Africa (green circle). From other articles connected to monkeypox, which we read when we were analyzing Figure 2, we have concluded that the Los Angeles monkeypox outbreak must have an African connection, so we hypothesize that the Assan Circus might lead us to the source of the virus and add it to our query. However, the resulting graph does not really tell us anything useful and we want to increase our neighborhood. Unfortunately, by this point the query with all of the various additions is returning a fairly large graph that would be unintelligible with a neighborhood of two. So instead, we do an isolated query on just the Assan Circus and graph it with a neighborhood of two. The result is shown in Figure 8.

Figure 8: Query results for Assan Circus with a neighborhood of two.

In Figure 8, the Assan Circus (middle pink circle) appears connecting two articles. The one on the right is the article that we have already read, which describes Vedric and Carnes’ (both circled in pink) efforts to track the circus and stop abuses. The article on the left, Week-of-Mon-20040301-1.txt_75, connects the circus to someone named Abu Hassan and references Zimbabwe. We then return back to our former query thread (last seen in Figure 7) and add these two new entities to see if we can expand our graph of relationships into Africa. The result of this query is graphed with a neighborhood of one in Figure 9.

Figure 9: Zoom in on query results, graphed with a neighborhood of one, after adding Abu Hassan and Zimbabwe.

Figure 9 shows that Abu Hassan and Zimbabwe (two middle pink circles) are referenced in the Import Permits spreadsheet and that he is connected to an animal import deal with a company called Global Ways (the column of pink circles on the left side of the graph). Viewing the Import Permits spreadsheet, we discover that Abu Hassan is the consignee for every import permit associated with Global Ways. We also see that the consignee address for Abu Hassan is different for each permit, with addresses in Tanzania, Zimbabwe, Zambia, Uganda, and Kenya. None of the animals imported into the US look like they would be carriers of monkeypox, except for perhaps the chimpanzee. However, it is from Zambia and not the Congo, so this thread is at a dead end for providing us with the source of the virus. However, the text of the article Week-of-Mon-20031215-1.txt_91 (green circle) provides further evidence that the Assan Circus is smuggling animals, so this is certainly part of an illegal smuggling network for exotic animals. This looks like a second subplot (the AJL activities in Los Angeles being the first) that has some connection to our main monkeypox plot through Carnes and Vedric. Because we want to explore as many connections to Global Ways as possible, we query on Global Ways in isolation, then graph it for a neighborhood of two. The results are shown in Figure 10.

Figure 10: Closeup of query results for Global Ways graphed with a neighborhood of two.

The import permits that Global Ways is involved in are mostly outside the graph window in the upper right corner of the view. These provide a graphical representation of the various permits with their associated animals and countries that were discussed above. The more interesting part of the graph is below the Global Ways node (top pink circle). The three articles circled in green all relate to the drug smuggling sub-plot. Miami and a phrase with ‘tropical fish’ (two left pink circles) graphically show the location and hint at the tropical fish connection. ‘South America’ is linked to all three articles and upon reading Week-of-Mon-20040105-1.txt_58, we discover the contaminated packaging and the symptoms of exposure (green circles in the text window on the right). Reading through all three articles provides all the details of this subplot.

This image also introduces two new people, Madhi Kim and r’Bear (blue circle), who are connected to Global Ways, Luella Vedric, and each other. We decide that the reference to r’Bert in Week-of-Mon-20040412-2.txt_13 is really a typo and was meant to reference r’Bear, so we use entity resolution to map resolve to r’Bear. This article describes the ‘Nights of Champagne and Tropical Fish’ auction that we list as one of our 20 events. It is significant because of the special guest status awarded by Kim to Vedric and r’Bear. The article Week-of-Mon-20040308.txt_109 (to the right of the blue circle) connects Kim and r’Bear to Shraavana and provides another important event and significant location. Upon viewing the article, we learn about the day r’Bear met with Kim at the sanctuary, which suggests that r’Bear may be a buyer of smuggled wildlife from Assan Circus through Global Ways. We need to learn more about Kim and r’Bear along with their connections back to the current solution subgraph, so we add them to the query that was made in Figure 9.

Figure 11 Gambian rat reference.

We will stop with our exploration at this point, since the use of the tool should be clear by now. We were unable to definitively resolve our two loose ends, the source of the monkeypox and how the chinchillas were infected. Since rodents are carriers, we tried a dead-ended query thread on laboratory animal abductions and lab break-ins, which we will skip over here. Then we found the comment about Gambian rats (green circle in text window in Figure 11) in the article Week-o-Mon-20040705.txt_86 (green circle in the graph view in Figure 11). The reference to the Gambian rats mysteriously dying and the fact that Ollesen is a business owner in Westminster, California, which is near Los Angeles, provides a hint that maybe they were the source. Explaining how the chinchillas were infected involved a similar logical leap based on Gil’s knowledge about Baptista and the question of who took the photographs. Perhaps Gil himself is the photographer, though without travel records to include in the database, we have no way of checking.

TOC: Who – What – Where – Debriefing - Process - Video