USC School of Cinematic Arts
Los Angeles, CA 90089
mpeteu ª usc ° edu
The relationship between geography, history and culture in the vast urban region of Los Angeles is explored through a practical, acessible, and expandable chrono-spatial visualization of its historic landmarks. CityTagz provides a means for residents of various ages to annotate and define their own communities through the identification of nearby resources. By uploading original media, proividing spatial data, and contributing descriptive tags, participants provide a multidimensional dataset that can be queried and rearranged however they see fit.
Figure 1 - Geographic annotation using Google API
Key words
Collaborative resource mapping, geographic annotation, geotagging, spatial cognition, bionic software, mental maps, shared spaces.
1. INTRODUCTION
Textile factories become apartment complexes, movie theatres are used for church services, and decrepit newspaper office buildings are transformed into movie sets. For a city quick to erase its own past, CityTagz attempts to decipher and pass on its fragile identity.
Los Angeles is a city of many unofficial boroughs [1]. Depending on whom you ask, some may even omit some of the municipalities based on personal preference, bias, or sheer lack of knowledge. CityTagz attempts to assemble a visual narrative of this conurbation by studying the individual history of each sector.
Since there are a multitude of placemark types that a landmark can be associated with, a flexible and consistent way to tag them must be introduced. Whether a user is adding information on Alvarado Street, the Chavez Ravine, the Civic Center or a renowned mural space, the experience will be similar.
The cartographic focus is not always on exact latitudinal and longitudinal coordinates, sometimes the information is related to a street name or an entire community. Therefore, a new way of tagging geographic landmarks has to be introduced. This project’s main objectives are resource annotation, accessibility, and spatial awareness.
Perhaps you have heard of a specific Los Angeles locale, and are eager to learn more about it. With CityTagz, you can either invest the time to travel to that place, gather media related to it, do some online reasearch, and post a short blurb of your own. Or, you can simply tag that spot on the map for someone else to fill in the blanks later.
The most agreed-upon tags offer the best description of a certain landmark. Since new tags are continuously added by new participants, this serves as a means to potentially change the portrayal of a particular place.
Visualization of the data is another important aspect of CityTagz. A strong visual organization contributes to the impact of the dataset on the viewer.
2. PRIOR ART
A basic example of a user-maintained map of global resources is Wikimapia [7], which attempts to break down large scale maps into subsections before allowing the user to post new data. Placeopedia takes this a step further by attempting to link every possible location on the globe with its related Wikipedia article.
The Green Map System [9] charts the natural and cultural environment in various cities across the globe while focusing on local interests. Another such example is a local PBS-sponsored project, Rites of Passage [6], exhibits Los Angeles area maps through children’s eyes. The concerns and interests specific to this age group makes for a unique viewpoint. The freedom given to the participants allows for the creation of highly personalized mental maps of their neighborhoods.
Following this trend of empowering locals to illustrate their community, London The Way We See It [17] is a photoblog which encourages participants to collaboratively map areas of they city by submitting pictures and short descriptions of each place. Unfortunately, there is no visualization of this data, therefore there is no way of quantifying the volume of posts, where they originated from, or what areas of the city are most frequented. This project takes advantage of technological permeation in today’s society, as more of us own digital imaging devices and have equal access to information.
Figure 2 - Norman Klein’s Bleeding Through
Norman Klein’s Bleeding Through: Layers of Los Angeles [12], recounts real-life stories that took place in the Downtown district in an elegant manner – the main fictional narrative is intertwined with seemingly congruent factoids related to Los Angeles history and culture. CityTagz emulates this style of storytelling by allowing individual participants to paint each locale based on their experiences.
Multiple well-established geovisualization tools exist: the Google and Yahoo Maps APIs (see Figure 1), which are web-based, and Google Earth and MSN Virtual Earth, which require the download of an installation package. Though their original intent was to facilitate transit, they have evolved into versatile “bring-your-own-dataset” authoring environments for data visualizations ranging from bike and run trails to mapping the origins of the 9/11 highjackers.
The ESP Game [10] is a web-based 2-player experience disguised as a game which cleverly uses “human computing cycles” to accurately describe all the images found on the internet. Now known as the Google Image Labeler, this creation of CMU professor Luis von Ahn has successfully engaged users for prolonged periods of time. According to his research, 9 billion human hours were spent playing solitaire in 2003. Von Ahn attempts to reallocate these hours to a more constructive task such as descriptive image tagging, a very complex problem for today’s computers. The ESP game takes advantage of human observation skills and stores user input which eventually benefits anyone who conducts an image search. This human-computer symbiosis is at the heart of CityTagz’s reward system. While categorizing existing resources in their own neighborhood, participants not only become co-authors of local history, but also benefit themselves and others seeking to better their knowledge of Los Angeles’s landmarks.
3. PROJECT DESCRIPTION
3.1 Scope
Traditionally, maps have had a limited amount of authors, and once printed, they become permanent. Maps suffer from these drawbacks, reinforcing the notion that maps are historical fiction seen from the “eye of the beholder”. CityTagz embraces that inherent bias, and unabashedly allows participants to experience each other’s skewed visualization of their locale and surroundings. Similar to many online services which allow for restaurant or venue reviews, they augment existing historical data with their personal accounts. CityTagz is used by students of various Los Angeles schools as well as members of the public. Anyone willing to create a free account and contribute data becomes an active participant in the mapping process.
CityTagz is intended to motivate people to visit landmarks in various areas of the city and ensure that public spaces are used to the full extent intended by their creators. The aim of this project is not to represent the street level dataset at the highest possible resolution, but transform the information into something more palatable. Ideally, participants would use CityTagz to find something that interests them using the clean interface, then find the exact cartographic location using a separate online service such as Google Maps.
The goal is to motivate participants to become return visitors and continue providing valuable data. If the feeling of ownership provided by the creation of a new landmark is not enough, then surely a tangible reward similar to Amazon’s Mechanical Turk [19] model will pique the participants’ interest. While Mechanical Turk provides a small sum for each “Human Intelligence Task” completed, CityTagz plans to reward the most active participants with gift certificates.
3.2 User Experience
At its inception, CityTagz was advertised to interconnected classroom units, yet it is available to any interested city inhabitant.
3.2.1 View Mode
Participants will be able to browse the entire map or select a specific borough (see Figure 3) and use the search capabilities in the query bar. More specifically, they will be able to show or hide all available layers, pan and zoom the map, and view the detailed information on each placemark. Also, users will be able to use the Timeshift slider and watch the landmarks appear, disappear, or change properties over the decades.
3.2.1 Edit Mode
Participants must be logged in order to take advantage of this mode’s add and edit features. This mode features the same actions that are possible in View Mode, plus the ability to add or edit your own posted placemarks.
The minimum required of each post is a landmark or resource name and its geographical location. In addition, participants may contribute a full text description of the place and the services it provides for its surroundings, as well as a media file (audio, video, still image). For classification purposes, multiple tags from a pre-orderdained list can be selected, and a related chronlogical table of events can be submitted.
Figure 3 - LA Municipalities selection screen
3.3 Data Gathering
An outreach program catered to existing geography or social studies classes in local public, charter, and private schools provides scores of enthusiastic participants who will populate CityTagz with rich personal accounts of routinely-utilized urban resources as well as locales frequented by each student participant. Each series of individual accounts could serve as an emergent narrative of their daily or weekly routine.
Digital still cameras and memory cards will be distributed on a classroom basis, and an instruction pamphlet on how to use cityTagz will be given to the students and instructor. Each participating classroom receives a hands-on CityTagz tutorial before the data gathering process begins. Student participants gather pictures of local resources and upload them from any computer with an internet connection, while providing precious metadata such as resource type, physical location, and weekday and/or time of day this resource is most heavily exploited.
Each resource type is associated with a descriptive icon, and its relative size is affected by the total amount of tags it has received. The largest icons will be clustered toward the center, with lesser-utilized resources radiating outward. Therefore, CityTagz is not meant to be a locative or reference map; rather, its primary purpose is to represent the epicenter of human activity in each borough. If necessary, any user can access a direct Google Maps link to the actual geographical location of each tag from the detailed resource view.
3.4 Visualization
The user interface consists of the actual map, pan and zoom controls, and the query bar. The map itself is broken down into four main GIS layers: ground level, ecology type, transportation network, and historical landmarks. The ground level and ecology types will be preset, while the transportation and landmark layer will remain editable. The map layers are each broken down into quantized hexagons, akin to the layout of a turn-based strategy board game. This pixel art-inspired surface will represent the break with traditional geovisualizations that increase in resolution with each zoom level.
CityTagz is presented in Flash, a vector-based authoring environment. Ironically, CityTagz will not take advantage of the accuracy of vector shapes; instead it attempts to simplify the existing landscape and provide a more palatable experience from a visual orientation standpoint.
3.5 Technical Overview
A PHP-based login system will allow users to browse as well as contribute to a condensed, semi-cartographic map of Los Angeles boroughs. A MySQL database stores all data submitted by participants. New posts consist of the following:
(1) resource name
(2) geographical location - the intersection of two streets and a city name
(3) one of multiple preexisting resource categories (entertainment, recreational, market, restaurant, health)
(4) media segment – a still image, an audio or video clip
(5) related historical timeline, where each event is associated with a specific year
This will be made possible through a Flash user interface which allows posting and viewing of various media. The system will first check against duplicate entries, and if a certain resource has already been added, then the user will be allowed to contribute their tag in addition to the existing entry. CityTagz purposely avoids the use of artificial language markers such as zip codes, latitudinal and longitudinal values, or mileage markers. The more tags a participant adds in a certain area, the more inflated it will seem compared to the surroundings. The participant can directly experience the personal map of someone residing in an entirely different location, and observe the places that have been tagged by that individual (or an entire group of locals). The system allows users to browse an overall map of Los Angeles, tags belonging to all users included, with the option to filter them based on category. The size of the corresponding icon for each represented resource will once again be determined by the number of tagz that particular location has received.
3.6 Potential Appendages
Still Imagery
Participants could peruse the comprehensive Los Angeles Public Library Photo Collection and attempt to match the exact angles of shots which reveal the transformation of assorted vistas between time periods.
Interest Groups
With the installment of a forum-style posting board, participants residing in different parts of the city could share information about where a certain resource is made accessible in their
Tag filtering
A collaborative filtering and tag ranking system would evolve CityTagz into an even more user-friendly and efficient place to quickly research available resources in Los Angeles. Also, user-defined associations between incongruent resource types could affect the clustering of placemarks in a new and unexpected fashions.
4. ACKNOWLEDGMENTS
My advisers, Jennifer Stein, Jennifer Swift, Perry Hoberman, Michael Naimark. Mar Elepano and Michael Naimark for providing LAUSD contacts. Special thanks to my ACM SIGCHI for providing conference-style paper templates.
5. REFERENCES
[1] Starr, K. A Borough System for Los Angeles. 18th annual Bollens-Ries Memorial Lecture. May 7 at UCLA
[2] Banham, R. Los Angeles: The Architecture of Four Ecologies. University of California Press, Berkeley, CA, 1971.
[3] Harmon, K. You are here: personal geographies and other maps of the imagination. Princeton Architectural Press, New York, NY, 2004.
[4] Wood, D. Power of Maps. The Guilford Press, New York, NY, 1992.
[5] Berry, K. and Henderson, M. Geographical Identities of Ethnic America. University of Nevada Press, Reno, NV, 2002.
[6] PBS – KCET. Rites of Passage. http://www.kcet.org/explore-ca/california-stories/ritesofpassage/maps/index.php
[7] Wikimapia. http://www.wikimapia.org
[8] A Chinatown Banquet. A multimedia history of Chinatown http://www.chinatownbanquet.org/about.html
[9] Green Map System, http://www.greenmap.com. Retrieved November 12 2006.
[10] O’Reilly, T. Google Image Labeler, the ESP Game, and Human-Computer Symbiosis. Retrieved Feb 18, 2007 from http://radar.oreilly.com/archives/2006/09/more_on_google_image_labeler.html
[11] USC – IMSC. GeoDec. http://infolab.usc.edu/projects/geodec/index.jsp
[12] Klein, N. and Kratky, A. Bleeding Through: Layers of Los Angeles. http://www.annenberg.edu/labyrinth/klein/klein.html
[13] Invisible Shape of Things Past. Film in virtual space. http://www.medienkunstnetz.de/works/invisible-shape/images/4/
[14] California Eagle Newspaper Photograph Collection, late 1800s-1950s. http://findaid.oac.cdlib.org/findaid/ark:/13030/tf6199n818
[15] The Center for Land Use Interpretation. http://clui.org/
[16] Los Angeles Public Library Photo Database. http://www.lapl.org/
[17] London, The Way We See It. Retrieved Feb 14, 2007 from http://london.thewayweseeit.org/
[18] Google Earth. http://earth.google.com
[19] Amazon Mechanical Turk. http://www.mturk.com