the data of visualization · 2019. 1. 28. · information visualization data as … • the usual...

11
INFORMATION VISUALIZATION The Data of Visualization

Upload: others

Post on 15-Feb-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

  • INFORMATION VISUALIZATION

    The Data of Visualization

  • INFORMATION VISUALIZATION

    Data as … • The usual sense of “data” is either a number, or a string of letters, dates, or as computer files. Let’s

    review/consider data …

    • booleans (true / false, 0 / 1)

    • integers (counting numbers, 1, 2, 3, …)

    • float (numbers with a floating point, e.g., 3.14159) [and there are other kinds of floats in computing, called “double”, “long”]

    • aggregates of data (e.g., x * y)

    • alphanumerics (characters, A-Z, 0-9)

    • bytes: characters (a single byte) or characters represented by their binary (0, 1), hexadecimal (base 16), octal (base 8) or similar ideas

    • nominal/categorial data are usually strings (e.g., “good|bad|average”)

    • ordinal, interval, ratio, as we find statistics …

    • Collectively these help us determine what visualization to use and, of course, are stored in computer files. The type of data and the type of file must be considered when ingesting the data for a visualization.

    !2

  • INFORMATION VISUALIZATION

    Data types and vis• Some data types suggest strongly what kind of representation to

    use; compare a table versus a chart. Compare these ways of expressing date data (e.g., in YYYY/MM/DD format).

    !3

    If you’re plotting data by a time frame, notice the different ways we can visualize them. Discuss. What kinds of audience, data, and tasks

    can be addressed by each of them?

  • INFORMATION VISUALIZATION

    Working with your data

    !4

    Scenario: You’ve been given a data file to visualize.

    •Check the data - are there missing values? 0 or null values? •Are the data in realistic range? [For instance, are there

    mistakes, say, 1,000,000 when the top range of the data is

    around 1,000?]

    •Check out the text, too. Should you convert the text to all capitals (easier to compare)? Are the data in UTF-8 encoding?

  • source data: cleaned of values outside domain and range; null values,inappropriate 0 values missing text characters; convert case

    most commonly: export the data as comma-delimited or tab-delimited text file

    script or program to convert the clean data to .json

    visualize!

    Typical real-world work scenario

    1 2

    3

    4

  • Geographic data: example of domain-specific data types (lat, long), combining categorical data, date and time, and numeric values

    !6

  • INFORMATION VISUALIZATION

    Models to engage usersBesides integrating and balancing data + visual languages, there is the story to be told … and as in any language reflecting on the approach helps to create the visualization as much as it does to help interpret it:

    • Descriptive

    • Persuasive

    • Explanatory

    • Exploratory

    • Process or Predictive

    • Data layout; Spatial layout

    !7

  • INFORMATION VISUALIZATION

    Data lifecycle• Start with the Needs or Requirements Analysis create a logical model of the

    system and then

    • The physical model [the programming and implementation]

    • Data Acquisition

    • Processing (data cleaning)

    • Analysis & Data or Statistical Modeling

    • Layout (Composition, Typography, Color, Data)

    • Reference System (part/whole relationships; grids; keys, etc.)

    !8

  • INFORMATION VISUALIZATION

    Data lifecycle• Projection/Distortion of the image (adjustments)

    • Access to source data by the end-user

    • Review the overall graphic design (ensure a harmony of the data/design for the audience, data, and task)

    • Aggregation & Clustering, redux: allow details on-demand, clustering, expanding data thru interactive techniques, such as zooming, data brushing)

    • Test with the End-Users: do the user-controls create a meaningful, interpretable story

    • Deploy!

    !9

  • INFORMATION VISUALIZATION

    Your own skills … On the basis of your experiences with HTML, CSS, JavaScript, and d3.js, encourage in-house testing of integrating local resources, such as local digital library collections, converting tab-separated value file (such as from a spreadsheet program) to .xml.

    !10

    For this example the exported tab-delimited data consists of a reference number, location, name, age, other fields. file name is tsvToXml.php

  • INFORMATION VISUALIZATION

    Even easier …There are a number of online conversion tools and a script to demonstrate the conversion.

    • Python and other languages make it even easier for anyone, including non-programmers, to convert a comma-separated value file into a .json file. In the Unix terminal window, run this line of code at the $ prompt, substitution ‘demofile.csv’ with your own file name:

    !11

    python -c “import csv,json; print json.dumps(list(csv.reader(open(‘demofile.csv’))))”

    https://www.commandlinefu.com/commands/view/20528/convert-csv-to-json