Create the center of typhoon information, gathering information from many eyes

Concept

The typhoon is a global-scale meteorological phenomenon. To capture such a global-scale phenomenon both in a macroscopic scale and in a microscopic scale, it is important to have the record of the phenomenon through a lot of eyes that see it. Many events occur simultaneously at many places as the typhoon makes a move. To know about what really happened during this period, it is necessary to have the collection of reports from the general public describing personal experiences.

The research project "Digital Typhoon" concentrated on creating the comprehensive collection of typhoon-related information from various sources. The following table represents the classification of various information in direct / indirect, public / personal dimensions.

Direct (Primary) Information Indirect (Secondary) Information
Public Satellite Images / AMeDAS Data Typhoon News Topics
Personal Eyephoon / Typhoon Front Typhoon News Weblog

Here direct information is produced without the process of aggregation and edition, while indirect information needs the process of aggregation and edition of direct information. In addition, personal information is always aware of the context of personal experience, while public information tries to exclude personality and focuses on more objective and global viewpoints instead.

This site "Eyephoon" takes the part of direct + personal information in the above table. That is, the target is the record of raw information directly published on the site describing personal experiences that might be subjective. This kind of information is micro and live in the sense that only a person who happens to be on the site of the typhoon can have an experience. The accumulation of this personal collection of information may lead to the digital archive of the typhoon which will be useful as the record of disasters.

Publishing, Aggregation, Organization and Dissemination in Participatory Information System

Valuable information of this kind is, on the other hand, difficult to deal with in the conventional system of journalism in the mass media. In the processes of aggregation and editing of information in the mass media, personal information is too microscopic and simultaneous, or too miscellaneous information as the elements of a story for a news article or program. If they make up a story from "picturesque" information, then we see standard information such as "waves that break on the wharf," which is often used in a TV broadcasting program on the spot. This kind of information may be picturesque but less real, probably because it lacks the transmission of information from the person who is involved in the event.

This is where participatory information system for publishing, aggregation and delivery can take part in. Information from the person involved is naturally of microscopic viewpoint, but really vigorous at the same time. Moreover, individual information may be meaningful only for a particular group of people, so we need a special mechanism to provide access based on information organization using geographic information, for example. Finally, we may be able to reconstruct global perspectives of the event from local perspectives by means of aggregating and editing miscellaneous information. This kind of massively-parallel collection of information may not be possible by the activity of the limited number of journalists.

The purpose of this site is to build a system to connect th spot and outside. Disaster information systems designed to work on the spot are getting more popular in national and local governments after the Hanshin-Awaji earthquake in 1995. But the main purpose of those systems is usually on the rapid understanding of and quick recovery from the disasters, and personal information is out of concern. The mass media, on the other hand, ranks their news for their needs, and that does not match well with personal information needs. Hence we aim to build a participatory information system for the typhoon and try to find best practices. In summary, the key is to find a method for collecting reliable information from undefined and large network of people using the infrastructure of the Internet.

Location-Embedded Trackbacks for Information Aggregation

Various mechanisms can be proposed for collecting information seen by massively-parallel "eyes" but on this site we first propose to use trackbacks from weblogs. One reason is in the wide popularity of the trackback as a tool to create links automatically, but more important reason is that the following characteristics of weblogs are suited as the sources of typhoon information.

  1. The system of "weblog + trackback" aggregates all information published by a person in the web space that is edited and maintained by the person. This system is good for accumulating information published from the person. If you use a bulletin board system, for example, information transmitted from the person is scattered across multiple bulletin boards, and information transmitted from the person is invisible from a historic perspective.
  2. The person is likely to have a motivation to keep the consistency of the web space in the course of editing material in that web space. This does not mean automatically that the contents of the web space is "correct" in any sense, but we can expect that the person feels a kind of "responsibility" in the editing process within the personal interest.
  3. We can understand the context of information published by the person by browsing past entries of the weblog. This is especially important in terms of the credibility of information, which is essential to deal with serious information such as typhoon disaster information.

Location information and temporal information is also regarded as relevant information on this site. A typhoon is a global-scale phenomenon, but specific events are usually local, and quite different from place to place. Hence location information is a key to get a better summary of the situation. Temporal information is also relevant because the situation changes quite rapidly on urgent situations.

Because of those reasons, we use "location-embedded trackbacks" for collecting location information using trackbacks. This convention describes location information by embedding location identifiers into the trackback URL. This may not be an intuitively easy way, but the advantage of this convention is that we do not need to modify existing widely-accepted systems, and this convention is seeing more and more usage for practical applications.

We use the followng identifiers for representing location information. Dividing the space of identifiers by the type and length of characters, we can omit a description about the meaning of information, which leads to the construction of concise Trackback URLs. This method cannot be universally applied, but in this system, we can find a systematic way for dividing the space of identifiers as follows.

Regular Expression Meaning
\w{2} ISO-3166-1 Country Codes
\d{2} JIS-X0401 Prefecture Identification Code (Japan)
\d{5} JIS-X0402 Towns and Villages Code (Japan)
\d{6} Typhoon Number YYYYNN
\d{7} 7-digit Zip (Postal) Code (Japan)
\d{8} Date YYYYMMDD

Weblogs and Moblogs

The name of the site "Eyephoon (previous name: Eyes on the Typhoon)" is chosen in the hope that we use not only "human" eyes but also more accurate "machinery" eyes such as digital cameras or cameraphones, for example, to record the world. Locator-equipped digital cameras and cell phones are also gaining popularity and can serve as mobile terminals with accurate location information. Those images may be especially useful to tell the the atmosphere of the spot for people outside of the spot.

The information fragment of images and text with accurate location information is already realized in moblogs and the like, and the real-time aggregation of those fragments can realize world-wide on-the-spot information delivery. Such massively parallel aggregation of on-the-spot information may lead to the unprecedented experience of a storm having a global scale.

Publications

  1. Asanobu KITAMOTO, "Typhoon Front: Spation-Tempral Interaction under Symbolic Large-Scale Natural Events", Interaction 2008, pp. 77-78, (in Japanese) [ Abstract ]

  2. Asanobu KITAMOTO, "URI Immunization: Eluding Spams in Participatory Systems", IEICE (Institute of Electronics, Information, and Communication Engineers) SIG of Web Intelligence and Interaction, No. WI2-2006-73, pp. 45-50, (in Japanese) [ Abstract ]

  3. Asanobu KITAMOTO, ""Digital Typhoon" and "Typhoon Front" - Typhoon Information Sites that Integrate Typhoon Data of the Past and Present", OHM, Vol. 93, No. 10, pp. 6-7, (in Japanese)

  4. Asanobu KITAMOTO, "URI Immunization to Elude Web Spam", NII Technical Report, No. NII-2006-010J, pp. 1-12, (in Japanese) [ Abstract ]

  5. Asanobu KITAMOTO, "Content Management Systems for Information Aggregation on Emergency Events such as Natural Disasters", The 19th Annual Conference of the Japanese Society for Artificial Intelligence, No. 3C3-02, (in Japanese) [ Abstract ]

  6. Asanobu KITAMOTO, "Digital Typhoon: Near Real-Time Aggregation, Recombination and Delivery of Typhoon-Related Information", Proceedings of the 4th International Symposium on Digital Earth, pp. (CD-ROM), [ Abstract ]

Acknowledgment

  1. This work is partially supported by The Okawa Foundation for Information and Telecommunications, "Database Systems for Collaborative and Instantaneous Organization of Disaster Information."
  2. We thank Japan Post for distributing the data file of zip codes (in Japanese), which is used for constructing the dictionary of trackback URLs.
  3. We thank University of Tokyo for maintaining CSV geocoding service, which is used for acquiring the location of municipal governments.