Visualizing Patterns
Finding hidden clues in Pittsburgh
10/31 Introduction to data visualization
We’re introduced to the project prompt this week and briefly did one exercise on thinking about data regarding water quality.

2 major purposes for data visualization:
- For us to better understand the hidden fact and pattern
- To communicate outwardly with others about our findings.
Keep in mind in this project
- use 3–5 types of data
- 20–30 points (not over 50)
- be remind of apples to apple comparison, cannot compare data sets that don’t have the same metrics
- Static data (structural/ geographical/ neighborhoods /land use/ area/ demographic)
- Dynamic data (Emergency/ Events changing)
Class brainstorm — Narratives around FOOD
Organic footprint in Pittsburgh (length of transportation form the source Farmer’s market/ Giant chain stores):
Food waste: (locations of giant chain stores & restaurant, food import and food consumption & rates of throw out).
Food source: How are farmer’s market soucing food from farmer’s farm. How are restaurants sourcing food from a farmer’s market
Accessibility: How is health food accessible to the whole Pittsburgh community?
Food Culture:( location, categories, residence demographics, prices)
Online Ordering habits: Who works at the latest every day? What kind of people orders online food? what are their professions? (time ordering food, frequency, location, profession, income, price, type, range)
Circular Economy in Food (food price/institutions)
The secret to a successful restaurant on Yelp: (name/price/location/length and diversity of the menu)
Bio-engineered foods relationship with health
A closer look at the Blooms-field Farmer’s market: (different types of organic/ the business model they have, the distance they travel to come here, the distribution of their food.)
11/07
Reading1: Chp 3. Representing Data
Different ways to configure information
- Cartesian / Polar/ Geographical(map)

Reading 2:
LATCH: Location. Alphabet, Time, Categories, Hierarchy. (Parts to a whole*)*Talk about the parts of a whole.
- Structure and organization of information help readers to extract value and significance from it.
Some tools for data vis:
- Datarwrapper: https://www.datawrapper.de/
- Rawgraphs: https://rawgraphs.io/
- d3 Javascript library: https://d3js.org/
Topics Forming (abandoned):
Farmer’s market: Where do they come from and who in Pittsburgh get access to? the locations of farmer’s markets and their relationship with gentrification and household income. One specific farmer’s market- Bloomfield and where does their vendor come from?
Online Ordering habits: Who works at the latest every day? What kind of people orders online food? what are their professions? (time ordering food, frequency, location, profession, income, price, type, range)
Tips for question framing:
- Note to mark data which is already there and data that is easy to find.
- If you couldn’t find the data you need, would your topic fall-apart? If so, jump shift to more suitable questions.
- What’s the order of information you want to go into?
Data Viz That interests me:
- Show different combination through different criteria


11/12 Sorting and support the Narrative with data
- apples to apples, can have subsets, try Separate Spreadsheets.
The relationship between Weather and Crime
Scale:
- Logarithmic (1, 10, 100)
- Linear (0, 1, 2, 3)
- Categorical (cloudy/sunny)
1. Determine the bucket of data
2. What might be the coordinate system for the data?
Things to do before Thursday
- Refine the questions
- Determin which scales am I working on
- Refine 3–5 buckets in each column(datasets)
- What is the coordinate system that can be used as an anchor
- create layers of information that we can find pattern and relations
11/14
What Types of Dataset have you collected?
- The active Alcohol License and Extended Permit dataset issued by PALCB (2019)
- I converted 36 Types of License into 4 Types of Catagory: (Manufacturing/Distribution/Consumption/Other use) (see left pic)
- After Grouping the Licensees by Zipcode, (see the blue and light blue chunks in the middle pic) I counted the quantity of each license category in every zip code area. (see right pic)
- I intend to use the data to depict the shape of the “infrastructure” for alcohol consumption in each zip code area.



Use zip code as coordinates, I mapped the Quantity of Consumption license, Quantity of Distribution license, Quantity of Importer license, and Quantity of other use as height in generating a spline, using Grasshopper, then the spline can revolve into a 3D shape:


2. Alcohol-related death (2018) Data from WPRDC
- Location(in zip-code) / Time / if happen with other OD / Date happened (if that matters)
- Demographic (underage drinking)
3. Alcohol-related crime (2018) Data from WPRDC
- 4 Categories: Sell liquor to minor / Disorderly Conduct/ Public Drunkenness/ Drunk drive
- Location/Neighbourhood
In-Class Exercise:
Research Question: What is the research question you aim to investigate through the data you collected?
Q: What are the alcohol consumption habits in Pittsburgh, and its relationship to crime,and death related with alcohol consumption?
Coordinate system: What coordinate system do you believe will serve as a good anchor for your piece?
- The Zipcode/Neighbourhood/Address (I have to locate into one coordinate)
Scales: For each type of data you plan to includes, what scale do you propose using?
Ranges: for each type of data and scales you established, what ranges/brackets/groups will you apply.
11/21:
Q: What question are you exploring?
Explore what kind of relationship there is between Alcohol-related death/crime and the distribution of alcohol industry across the Pittsburgh
Q: What types of data are you using?
- 4 categories of alcohol-related data (originally 36)
- The number of alcohol-related licenses issued in each zip code
- The number of alcohol-related death in the year 2018 in zipcode(both )
Q: What co-ordinate system are you using?
Using Geographical (zipcode) as a coordinate system, because the question explores the distribution status.
Q: What scales are you trying to do in each of the data?
categorical:
- different types of alcohol-license
- different types of alcohol-related crime
location(relative position)
- distribution in zipcode
Linear:
- quantity of alcohol license issued in each zip code
Q: What are the ranges in your data?
- categories of alcohol-related data (1,2,3,4)
- The number of alcohol-related licenses issued in each zip code
- The number of alcohol-related death in the year 2018 in zipcode(both )
Q: What would be the indexical stories and your narrative?
- (Category)what kind of alcohol license is issued in Pittsburgh?
- (Part of a whole)how do they form an industry? what is their relationship?
- (Location) What is the alcohol license condition in every zip code area?
- (Category/Linear)What is the quantity of each kind of license in every zip code area?
- (Linear) What is the quantity of alcohol-related death in every zip code?
- (location)8 people travel elsewhere after drinking and died.
- (Linear/Catagory) What is the quantity and categories of alcohol-related crime in every zip code?
Q: What visual or temporal cue are you using in the design?
- (Category)what kind of alcohol license is issued in Pittsburgh?
- (Part of a whole)how do they form an industry? what is their relationship?
- (Location) What is the alcohol license condition in every zip code area?
- (Category/Linear)What is the quantity of each kind of license in every zip code area?
- (Linear) What is the quantity of alcohol-related death in every zip code?
- (location)What are the outliers(incident zipcode → decedent zip code)?
- (Linear/Catagory) What is the quantity and categories of alcohol-related crime in every zip code?
The purpose of the class:
Building up Cognitive models is not to present the structure of information but the insight of data and information.
Looking into the interaction in the design
Pattern + Detection
- Bucketing helps detect the pattern in the information
- Categorization around 7 is the maximum of perception
- Temporal building (bring someone into the information)
Representation
- Categorization
- pacing + simultaneity. show relationships between relative data
- Narrative + indexical stories
- Expectations + Perception (Norman, D., Thing that makes us smart, Appropriateness principles)
I can use background color change in the narrative to represent different tones. Using the light color(wine/beer/vodka/Champagne) to introduce the industry, use desaturated color to introduce crime and death.
Interaction
- customization (provide people with control/how much choice)
- mimicking known behaviors (builds on what people know, but at the same time push the boundaries, bring something new)
Experience
- recall engagement. Setting the context
- discovery critical thinking (not include text in the beginning . /How do you engage the user to find the pattern?)
