What Makes Paris Look Like Paris?

  • strict warning: Only variables should be assigned by reference in /courses/i202/public_html/f12/includes/common.inc(1696) : eval()'d code on line 1.
  • strict warning: Only variables should be assigned by reference in /courses/i202/public_html/f12/includes/common.inc(1696) : eval()'d code on line 1.
  • strict warning: Non-static method view::load() should not be called statically in /courses/i202/public_html/f12/sites/all/modules/views/views.module on line 879.
  • strict warning: Declaration of views_handler_argument::init() should be compatible with views_handler::init(&$view, $options) in /courses/i202/public_html/f12/sites/all/modules/views/handlers/views_handler_argument.inc on line 745.
  • strict warning: Declaration of views_handler_filter::options_validate() should be compatible with views_handler::options_validate($form, &$form_state) in /courses/i202/public_html/f12/sites/all/modules/views/handlers/views_handler_filter.inc on line 589.
  • strict warning: Declaration of views_handler_filter::options_submit() should be compatible with views_handler::options_submit($form, &$form_state) in /courses/i202/public_html/f12/sites/all/modules/views/handlers/views_handler_filter.inc on line 589.
  • strict warning: Declaration of views_handler_filter_boolean_operator::value_validate() should be compatible with views_handler_filter::value_validate($form, &$form_state) in /courses/i202/public_html/f12/sites/all/modules/views/handlers/views_handler_filter_boolean_operator.inc on line 149.
  • strict warning: Declaration of views_plugin_row::options_validate() should be compatible with views_plugin::options_validate(&$form, &$form_state) in /courses/i202/public_html/f12/sites/all/modules/views/plugins/views_plugin_row.inc on line 135.
  • strict warning: Declaration of views_plugin_row::options_submit() should be compatible with views_plugin::options_submit(&$form, &$form_state) in /courses/i202/public_html/f12/sites/all/modules/views/plugins/views_plugin_row.inc on line 135.
  • strict warning: Only variables should be assigned by reference in /courses/i202/public_html/f12/sites/all/modules/social_media/social_media.module on line 248.

Data mining is the process of organizing and structuring random data so as to extract meaningful information from large datasets. This article by Emily Badger published on August 8, 2012 in The Atlantic Cities describes one such data analytics project from Carnegie Mellon University and INRIA/Ecole Normale Supérieure in Paris. These researchers have developed an algorithm to process large datasets of geotagged imagery with the goal of distilling minute stylistic elements in a city’s urbanscape that impart the city it’s distinct visual identity. For instance, if given several random images of streets and boulevards in Paris, the algorithm would extract visual cues such as the distinctive Parisian cast-iron balconies, and decorative street signs and lamp posts as features that are copious in Paris, but at the same time are unique only to that city.

The algorithm processes millions of patches of streetscape images from the Google Street View database, and filters and structures it using two organization criteria: frequency and distinction. Several visual elements in a city’s streetscape such as trees and cars occur frequently, but are not visually distinctive. Contrarily, an iconic structure like the Eiffel Tower is highly distinctive, but only one of it’s kind. However, many distinctive stylistic elements in the visual fabric of a city (like the Parisian balconies) occur repeatedly to form patterns that are different from patterns elsewhere, and serve as visual identifiers that are unique to that city. In comparing and sorting different patches of random images, the algorithm optimizes for both frequency and distinction to rank and extract these unique identifiers. Using GPS coordinates of each image, the occurrences of the identifiers can then be mapped to construct visual stylistic narratives that portray architectural interactions and influences within and across cities.

Urban studies are typically macroscopic and address the design and planning of urban areas at a very broad scale. The identity of a city is conventionally defined by it’s districts and neighborhoods, iconic buildings and skylines, streets and public spaces, and urban amenities. However, this research shows how something as microscopic as the “visual minutae of daily urban life” can be isolated and collectively mapped to represent distinctive visual identities within and across streets, neighborhoods, districts, and cities. These identities, shaped by influencers such as culture and values, historical and geographical context, economic vitality, etc. can then be used to understand various social and cultural interactions across different regions and time periods.