Austiknights

A web application for finding places in Austin

Transcript

1. Austiknights CS 373 - Software Engineering
2. The Team Fatimah Zohra Nick Crum Keerthana Kumar Safaa Hansen Yun Jang Frontend Backend Search Server Unit Testing Backend Data Retrieval Unit Testing Search Server Continuous Integration API Integration Frontend Unit Testing Server Continuous Integration Unit Testing Frontend Backend Continuous Integration Backend Data Retrieval Unit Testing
3. The Purpose • Give users a convenient way to get locations in Austin. • Gets relevant and popular results based on search. • Pillars: Places, Categories (Types), Reviews
4. Technologies • Google Places and Maps API • Frontend (AngularJS, Bootstrap) • Backend (Python, Flask) • Database (PostgreSQL) • Apache Solr • Travic CI
5. Data Mining Stored over 50 MB of mined data! Over 17,000 unique places in Austin!
6. Models & Database Design
7. Demo - Walkthrough APIary TYPES PLACES SEARCH DETAILS REVIEWS http://docs.austiknights.apiary.io/#
8. Demo - Walkthrough Site • Home • Places ◦ Details / Reviews • Maps • Cuisine • Test • About Us
9. Critiques - Us What did we do well? • Clear, content-focused UI • Rich data with plenty to explore • Integrating technologies for a more powerful product, i.e. Solr, Google Places, AngularJS.
10. Critiques - Us What did we learn? • It is important to isolate steps so that a full build isn’t necessary in order to make incremental changes. For example, data mining, migration to the database, frontend design and search can all be developed with only the occasional full build. • Write tests early lest they all fail hours before the deadline. • Important to understand the full development stack.
11. Critiques - Us What can we do better? • A better second story. What happens after the first search is conducted? • Reduce storage of data to one place (remove json middle man) • Improve the search algorithm
12. Critiques - Us What puzzles us? • How do you make objectively functional and elegant UI? What are the current best practices? • What does a user want from our site? • Best pattern matching for searching
13. Critiques - Food IDB What did they do well? - an attractive variety of embedded media - an intuitive and easily navigable UI - a well documented and selective RESTful API
14. Critiques - Food IDB Dothraki Blood Pie?! What did we learn from their website? - How three well defined pillars can be organized with equal importance
15. Critiques - Food IDB What can they do better? - Increase the data - Provide a unique service for the data (analysis tool)
16. Next Steps • Better text searching, less noise • Integrate featured deals for places • Ranking for hot places in Austin • Open hours for places • Pricing for places • Search by location
17. Next Steps • Updated data-mining algorithm