Deviant Login Shop  Join deviantART for FREE Take the Tour
List of Open Problems
- Find text that looks like an address in a given text
- Match hawker center reviews to its specific stalls
- Find known entities (from the db) in a given text (e.g. dish names, etc) -- useful to turn phrases in the text as hyperlinks

5th July 2012
Meeting #10

7 June 2012
Meeting #9
[Jesse ]
- Zhaoyan's review quality+sentiment code finished running on all review in db
- main page UI done; what data should we show?
- search results and  list (index) page UI; Yilu to design look
- Rewrote script to download and parse LIC articles; updated both (LIC and review) tables in db
- Gave up on processing insing articles. What we really need is the address and stall name. The rest is easy: review content, title, timestamp, author name, url. Zhaoyan to help with code to get address and stall name from an article
- TODO: [MOST IMPT] prepare script for demo, e.g. choose 2 popular hawker centers. 5minute demo. Get all the pictures and list of reviews from various sources and they have links to their original reviews.
- TODO: migrate to mysql
- TODO: integrate Xiangnan's code
- TODO: integrate Zhoayan's wikipedia code (dont show dish names that do not have matching wikipedia articles)
- TODO: send meeting reminder
- TODO: overlay image source (url) on top of its image
- TODO: think of review UI (separate different sentiments in the UI?)
- TODO: make map markers clickable
- Made code for finding image urls and source web page for a query of dish, stall, or hawker center
- [TODO] look at ieatishootipost data and see how we can extract the stall name from the articles. You could obtain a list of stall names from the sqlite3 database at next@han.ddns:/home/next/rails/sg.eats/db/development.sqlite3. Please copy the file to your own working directory.
- Made new logo designs
- Made new rating designs
- TODO: make more rating designs (e.g. food items for half star and full star) that works best at the small size in which it will be displayed.
- TODO: look at the top section of… (links to A, B, C, …, Z)  and propose a sketch/design of how these links should be displayed
- TODO: look at the top section of… ( the options to display hawker center, stall, etc) and propose a sketch/design of how these check box options should be displayed
- canonicalized dish names
- linked wikipedia entries to dish names
- ieatishootipost data fetched, but no link information available -- need to re-process
- TODO: look at how to get address from insing articles. Reprocess insing articles

10 May 2012
Meeting #8
[Jesse ]
- rank based on click count
- Edited stall page, dish page
- Image CSS shadows, other CSS changes based on chosen design by Yilu
- Added pagination for reviews
- Done with hawker center page
- Processed foursquare reviews
- Processed ladyironchef
- Processing insing
- Wikipedia article update
- TODO: fill meeting minutes
- TODO: set time for next month's meeting
- TODO: send a list of dishes <name, id> to Zhaoyan
- TODO: integrate Zhaoyan's review quality+sentiment code
- TODO: main page UI
- TODO: search results UI
- TODO: Allow sorting for list of dishes, hawker centers, stalls
- TODO: continue insing articles processing
- TODO: re-process ladyironchef text content from original url and replace in db
- Made 4 logo designs
- Not done with star rating designs
- TODO: make design permutations on the design with the yellow text with green outline to vary the position and size of the table graphic in the logo
- TODO: star rating design, preferably something unique; need not look like a star; preferably matches the design of the logo
- Java code to give a review a quality score and a sentiment score
- TODO: write code to canonicalize dish names in the db
- TODO: For each dish name in the db, give description of the dish from the first paragraph of its wikipedia article, if any.
- TODO: scrape articles (keep the raw html for each post as well)
- TODO: <Choose an open problem to solve>
- No update
- TODO: make code that given a query string and a type (stall, hawker center, dish) will produce a list of 5 image urls and their corresponding page urls using google image search; each type should use certain site-specific clauses (e.g. for dish use, for hawker center use, etc)

5 Apr 2012
Meeting #7
[Jesse ]
- Finished with mockups…
- Yilu made different color schemes…
- Changed db schema
- Removed blacklight
- friendly urls
- sunspot solr
- image gallery
- admin login, separate from user
- google maps
- Imported from Xiangnan's Foursquare table: stall, hawker centers, users, reviews, dishes
- rough clean up of stall names and dish names. needs more work
- populate hawker center lat lng, adapted from Zhaoyan's code. Now a rake task.
- same-page pagination via ajax
- Examples of:
Stall page:…
Dish page:…
Stall list:…
>> need more real data (and connections) to work with.
- TODO: Wikipedia API
- TODO: continue rails work, use green design from Yilu (but add closing quote)
- TODO: fill-up minutes for April and March meetings
- TODO: click-count-based ranking of models

[ Yilu ]
- Different color schemes and logo designs:…
- TODO: graphic for star ratings

[ Zhaoyan ]
- Downloaded data from ladyironchef and and foursquare reviews from Varun
- TODO: ascertain quality of each review (e.g. Fogg index, Fleisch - Kincade; readability score ask Zhao Jin)
- TODO: pass raw data for foursquare reviews to Jesse
- TODO: check number of hawker centers in google maps sample code
- TODO: download data from ieatishootipost

[ Xiangnan ] no updates

5 Mar 2012
Meeting #6
[ Jesse ]
- specs for each model view
- per-page Photoshop mockups
- logo proposal
- TODO: start on rails app
- TODO: finalize design with Yilu offline

[ Zhaoyan ]
- google map distance function sample page
- TODO: ascertain quality of review
- TODO: download blog data

[ Yilu]
- Fireworks mockups for dish, stall pages
- TODO: tag line
- TODO: make logo pretty
- TODO: work with Jesse with final design

[ Xiangnan ] no updates

9 Feb 2012
Meeting #5

This looks like Unix time.
1268927156 = Thu, 18 Mar 2010 15:45:56 GMT
And a code sample:
DateTime startDate = new DateTime(1970, 1, 1);
DateTime time = startDate.AddSeconds(1268927156 );

Xiangnan: downloaded Foursquare data from "102 Hawker_Center group".  Hawker centers match, but the stalls don't match

Todo: Min: Write Varun to crawl foursquare (Xiangnan URLs for 102 hawkers centers)
Todo: Xiangnan: Write a <WebSite>User table.  WebSiteUser_User join table (which has the join fields + WebSite_Type {e.g., User_ID, WebSiteUser_ID, WebSite_Type}
Todo: Zhaoyan: Add map feature to the show.html.erb for stalls and hawker_centers,
Todo: Min: Jesse to get admin rights to han
Todo: Min: review controller
Todo: Jesse: to migrate sg.eats to han.ddns  
Todo: Jesse: to make it look nice

5 Jan 2012
Meeting #4


Zhaoyan: 1) Map locations to GPS

     2) Heroku connection to Jetty
Xiangnan: 1) Foursquare (API or not) ingest some hawker center data, reviews

Entity Linking as subtask
Min:     1) Review model and its connectivity

2) Put a full RA on making it look nice

13 Dec 2011
Meeting #3

Min, Xiangnan, Zhaoyan   11:00-11:30am via Google+ Hangout

New Tasks for 13 Dec 2011- 27 Dec 2011

Xiangnan:  1) Work on the UI of sgEats project with CSS and javascript

           2) Try out partial and learn from and Yelp UI

Zhaoyan:  1) Continue working on indexing (with updating schemes) and searching
                 2) Modify the schema to include photo, using blob for multimedia data
                          3) Import Adam Road Hawker Center data

Min:          1) Create model for reviews

Min and Zhaoyan meet after two weeks; three meet in early January.

Finished Tasks

Xiangnan: 1) Canonized stall_categories (about 40 types) for data
                 2) Ingested stalls and stall_categories into the project

Zhaoyan: 1) Explored Blacklight for searching
                            2) Obtained Adam Road Hawker Center data from KAI Square,

including all the dishes and stalls of the HC and photos of the stalls.

28 Oct 2011
Meeting #2

- We decided to use ruby 1.9.2 and rails 3.1.1
- We decided to use heroku for deployment / and our local repository
- Three tasks:
1. UI - all to do this [ruby-toolkit]
2. Data Model - Min to work on this
3. Sample Data - Zhaoyan from, Xiangnan from NUS Arts Canteen