Openwave

Openwave

Problem:
  • Future-proof: The ability to reuse the system for other sports turned out to be very daunting. Replication became increasingly difficult as new features were added. We ended up take notes of where we coded tennis-specific solutions. Though an automated solution wasn’t available within the clients budget, the combination of the existing code and documentation would give clear direction, should Juump expand to additional sports.
  • Data Collection: Tennis court location information doesn’t exist in a central hub--and information can vary greatly between sites. We felt that the only reliable way to verify the validity of tennis court locations was to cross reference at least 2 different sources, and add another level of human checks. As a result, we implemented:
  • A spider code that scrapes websites for tennis information. This was coded in Python programming language, due to its excellent text processing capability and the ease of coding to allow flexible changes, as we encounter new data format from new data sources.
  • Automatic Cross referencing and court merging tool, also coded in Python.
  • An interface with Amazon Mechanical Turk (https://www.mturk.com/mturk/welcome), the work force on demand service provided by Amazon. We built a web form system to allow mturk workers to directly verify our data.
Challenge:
  • Drupal was chosen as the open source CMS framework. Out of the box, Drupal provides a lot of customizable content, so we can create events, groups, user profiles and tennis courts very easily. When it comes to implementing the queries to pull out search results, we hand code custom modules to achieve the speed required.
  • Future-proof: The ability to reuse the system for other sports turned out to be very daunting. Replication became increasingly difficult as new features were added. We ended up take notes of where we coded tennis-specific solutions. Though an automated solution wasn’t available within the clients budget, the combination of the existing code and documentation would give clear direction, should Juump expand to additional sports.
  • Data Collection: Tennis court location information doesn’t exist in a central hub--and information can vary greatly between sites. We felt that the only reliable way to verify the validity of tennis court locations was to cross reference at least 2 different sources, and add another level of human checks. As a result, we implemented:
  • A spider code that scrapes websites for tennis information. This was coded in Python programming language, due to its excellent text processing capability and the ease of coding to allow flexible changes, as we encounter new data format from new data sources.
  • Automatic Cross referencing and court merging tool, also coded in Python.
  • An interface with Amazon Mechanical Turk (https://www.mturk.com/mturk/welcome), the work force on demand service provided by Amazon. We built a web form system to allow mturk workers to directly verify our data.
Solution:
  • Google Maps API: We chose the Google Maps API as the underlying mechanism to show search results for tennis players, events, courts and groups. All “content” of the Juump system has a real latitude and longitude identifying its location, and corresponds to a map showing nearby tennis players, events, courts, and groups/clubs.
  • Drupal was chosen as the open source CMS framework. Out of the box, Drupal provides a lot of customizable content, so we can create events, groups, user profiles and tennis courts very easily. When it comes to implementing the queries to pull out search results, we hand code custom modules to achieve the speed required.
  • Future-proof: The ability to reuse the system for other sports turned out to be very daunting. Replication became increasingly difficult as new features were added. We ended up take notes of where we coded tennis-specific solutions. Though an automated solution wasn’t available within the clients budget, the combination of the existing code and documentation would give clear direction, should Juump expand to additional sports.
  • Data Collection: Tennis court location information doesn’t exist in a central hub--and information can vary greatly between sites. We felt that the only reliable way to verify the validity of tennis court locations was to cross reference at least 2 different sources, and add another level of human checks. As a result, we implemented:
  • A spider code that scrapes websites for tennis information. This was coded in Python programming language, due to its excellent text processing capability and the ease of coding to allow flexible changes, as we encounter new data format from new data sources.
  • Automatic Cross referencing and court merging tool, also coded in Python.
  • An interface with Amazon Mechanical Turk (https://www.mturk.com/mturk/welcome), the work force on demand service provided by Amazon. We built a web form system to allow mturk workers to directly verify our data.
Result:
  • Google Maps API: We chose the Google Maps API as the underlying mechanism to show search results for tennis players, events, courts and groups. All “content” of the Juump system has a real latitude and longitude identifying its location, and corresponds to a map showing nearby tennis players, events, courts, and groups/clubs.
  • Drupal was chosen as the open source CMS framework. Out of the box, Drupal provides a lot of customizable content, so we can create events, groups, user profiles and tennis courts very easily. When it comes to implementing the queries to pull out search results, we hand code custom modules to achieve the speed required.
  • Future-proof: The ability to reuse the system for other sports turned out to be very daunting. Replication became increasingly difficult as new features were added. We ended up take notes of where we coded tennis-specific solutions. Though an automated solution wasn’t available within the clients budget, the combination of the existing code and documentation would give clear direction, should Juump expand to additional sports.
  • Data Collection: Tennis court location information doesn’t exist in a central hub--and information can vary greatly between sites. We felt that the only reliable way to verify the validity of tennis court locations was to cross reference at least 2 different sources, and add another level of human checks. As a result, we implemented:
  • A spider code that scrapes websites for tennis information. This was coded in Python programming language, due to its excellent text processing capability and the ease of coding to allow flexible changes, as we encounter new data format from new data sources.
  • Automatic Cross referencing and court merging tool, also coded in Python.
  • An interface with Amazon Mechanical Turk (https://www.mturk.com/mturk/welcome), the work force on demand service provided by Amazon. We built a web form system to allow mturk workers to directly verify our data.
Copyright © 2008 - 2016 luxcer.com. All Rights Reserved.

Login

ERROR: Invalid username.