Major studio one

project collaboration w/ Urban Ecology Lab

Parsons Data Visualization (MS)

X

data sources

original dataset:

• .1. urban ecology lab - map PLUTO

example:

• .2. open data NYC - restaurant inspection data

example:

CAMIS	DBA	BORO	BUILDING	STREET	ZIPCODE	PHONE	CUISINE DESCRIPTION	INSPECTION DATE	ACTION	VIOLATION CODE	VIOLATION DESCRIPTION	CRITICAL FLAG	SCORE	GRADE	GRADE DATE	RECORD DATE	INSPECTION TYPE
50036304	NANOOSH	MANHATTAN	469	7TH AVE	10018	2123900563	Mediterranean	7/10/15	Violations were cited in the following area(s).	09C	Food contact surface not properly maintained.	Not Critical	7	A	7/10/15	2/27/17	Pre-permit (Operational) / Re-inspection

• .3. foursquare API

example: API endpoint: venue | API endpoint: id/search

• .4. dataset I can not found but need:

   median household income NYC (scale: per lot/block)
  

dataset after joining/parsing:

example:

name	address	boro	lat	lng	postcode	categories	stats	rating	text__type	text__text_extract	violation__recentScore	violation__recentTime	violation__historyScore	violation__historyVCode	violation__closed_before	id
Paris Baguette	136-17 39th Ave	MANHATTAN	40.76063451	-73.82956266	11354	Bakery	7172	7.9	liked	comfortable	10	2017-02-03T05:00:00.000Z	10	06C	null	0
										long			10	06E
										other			7	06C
													9	04K
													9	08A

topic

• combine Restaurant Inspection data and foursquare review/rating data, find interesting pattern between restaurant's sanitation level (inspection data) and what New Yorker's real reflection (foursquare API) about those restaurant. f.e: top 10 filthiest restaurants new yorker's love the most
• find the relationship between household income and restaurant sanitation problem. f.e: base on geolocation(each street block/district), mapping income and restaurant sanitation score in the map

key words:

Sentiment analysis: foursquare customer review text analysis. extract all Adjectives / Adverbs base on positive / negative context.

Manhattan: focuing all the restaurant in Mahattan(~ 5500 restaurants).

Filthy-but-lovin'it INDEX: Quantify the relationship between restaurant inspection score and foursquare rating (1-10).

Design Sketches

plan A:

mobile design sketches & third design sketches

prototype

welcome page/Landing:

Because All New Yorker's think NYC as "center of the universe", so I decide to make the main visualization shape as a planet, in the center, each big circle represent each violation categories(total 78); and all the small dots obits represents each restaurants in Manhattan.

• on the top-right, is a navigation area, to lead reader's explore the visualization graph.

• on the bottom-right area, all the text, location, sanitation problem and foursquare review text change programmatically when you mouse over small dots(which represents each restaurants in Manhattan)

• the location of the dots in this graph is assigned by force layout algorithm, so that is NOT mapping base on Lat,Lng.

demonstrate when you mouseover restaurants. (sanitation problems are both shows on the tooltips and bottom-right area) :

demonstrate when you mouseover/click(buttons in top-right) sanitation problems. (tooltips will show the total number of restaurants which have this sanitation problem. bottom-right area will not change text when you mouseover sannitation problems):

demonstrate when you click districts button. (tooltips will show the total number of restaurants which have this sanitation problem. bottom-right area will not change text when you mouseover sannitation problems) :

analysis page (second slide):

• use Filthy-but-lovin'it INDEX, to calculate top ten restaurants that is disgusting but ppl love to go.

• the three scatterplot in the page is mapping base on inspection score, restaurant grade(A,B,C) and foursquare rating score respectively.

demonstrate when you click top ten restaurant, it will show you detailed info about their inspection history within two years.

analysis page (third slide):

mapping each restaurant on the map with two variables: sanitation level and household income (per st. block) and write a paragraph or two to highlight the finding.

plan B:

prototype

Landing page:

Use map and to show each restaurants in Manhattan, and navigation button on the right hand side. because Manhattan is a narrow island, so I dicede use up and down direction to navigate though the map.

• use a lot of button to change the level / type of measurement
• color code base on level / type of measurement
• flyby animation when you click the nav button.
• tooltips show the detail info of each restaurants(sannitation problem, foursquare sentiment analysis result, etc).

analysis page (second slide):

This mock up shows the idea that the analysis page I want to use one chart show all the relationship between foursquare rating; inspection score, household income etc.(the form of chart haven't decided, but it definitely goona be one zoomable chart).

• the x axis is ordered by latitude(up-to-down).
• the y axis have lots of types/ level of measurement, that will show the relationship between each restaurant that ordered by geolocation(x axis).