02 Process
-
Upload
tusharnimbhorkar -
Category
Documents
-
view
233 -
download
13
description
Transcript of 02 Process
![Page 2: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/2.jpg)
This Week• HW0 - due next Tuesday (not graded)
• Install Anaconda & IPython frameworks
• Sign up for Piazza and introduce yourself
• Fill out survey
• Friday lab 10-11:30 am in MD G115
• Intro to Python with Ian Stokes-Rees
• Make sure to have IPython installed and ready
• Readings - post comments on Piazza
![Page 3: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/3.jpg)
Outline
• Process & Process Books
• What makes visualizations effective?
• Data Sources & Data Cleanup
![Page 4: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/4.jpg)
Process
![Page 5: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/5.jpg)
Data ExplorationNot always sure what we are looking for (until we find it)
![Page 6: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/6.jpg)
Ask an interesting question.
Get the data.
Explore the data.
Model the data.
Communicate and visualize the results.
What is the scientific goal?What would you do if you had all the data?What do you want to predict or estimate?
How were the data sampled?Which data are relevant?Are there privacy issues?
Plot the data.Are there anomalies?Are there patterns?
Build a model.Fit the model.
Validate the model.
What did we learn?Do the results make sense?
Can we tell a story?
![Page 7: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/7.jpg)
What do analysts do?
![Page 8: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/8.jpg)
What do analysts do?
![Page 9: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/9.jpg)
What do analysts do?
I spend more than half of my time integrating, cleansing and transforming data without doing any actual analysis. Most of the time I’m lucky if I get to do any analysis. Most of the time once you transform the data you just do an average... the insights can be scarily obvious. It’s fun when you get to do something somewhat analytical.
![Page 10: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/10.jpg)
“The greatest value of a picture is when it forces us to notice what we never expected to see.”
John Tukey
Exploratory Data Analysis
![Page 11: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/11.jpg)
Ascombe’s Quartet
Anscombe ’73
Same mean, variance, correlation, and linear regression line
![Page 12: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/12.jpg)
Ascombe’s QuartetSame mean, variance, correlation, and linear regression line
Anscombe ’73
![Page 13: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/13.jpg)
Example: AntibioticsWill Burtin, 1951
![Page 14: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/14.jpg)
Effectiveness of Antibiotics
![Page 15: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/15.jpg)
Data & Questions
• What are the data types?
• What are possible questions?
![Page 16: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/16.jpg)
Data• Genus & species of bacteria [string]
• Antibiotic name [string]
• Gram staining? [pos/neg]
• Minimum inhibitory concentration (mg/ml) [float] (lower == more effective)
![Page 17: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/17.jpg)
What Questions?
![Page 18: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/18.jpg)
M. Bostock, Protovisafter W. Burtin, 1951
How effective are the drugs?
P & N N
![Page 19: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/19.jpg)
Wainer & Lysen, “That’s funny...”American Scientist, 2009
Adapted from Brian Schmotzer
Not a streptococcus!(realized ~30 years later)
Really a streptococcus!(realized ~20 years later)
How do the bacteria compare?
![Page 20: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/20.jpg)
Wainer & Lysen, “That’s funny...”American Scientist, 2009
How do the bacteria compare?
![Page 21: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/21.jpg)
“The greatest value of a picture is when it forces us to notice what we never expected to see.”
John Tukey
![Page 22: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/22.jpg)
Process Books
![Page 23: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/23.jpg)
Process Books
[Varun Bansal, Cici Cao, Sofia Hou, CS171, 2013]
[Blake Walsh, Gabriel Trevino, Antony Bett, CS171, 2013]
![Page 24: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/24.jpg)
IPython Notebookshttp://nbviewer.ipython.org/
![Page 25: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/25.jpg)
IPython is Great(for large-scale computation, data exploration, and creating reproducible research artifacts)
Mike Roberts, Stanford University
[http://nbviewer.ipython.org/urls/raw.github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/master/Chapter1_Introduction/Chapter1_Introduction.ipynb]
![Page 27: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/27.jpg)
E.g.: https://github.com/jrjohansson/scientific-python-lectures
Intro to Python Lab this Friday!!
10-11:30 am, MD G115
![Page 28: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/28.jpg)
You probably all know the default Python interpreter.
Don’t bother with
![Page 29: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/29.jpg)
IPython is a more powerful interactive Python interpreter.
write and execute Python code in
snippets
[http://nbviewer.ipython.org/urls/raw.github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/master/Chapter1_Introduction/Chapter1_Introduction.ipynb]
![Page 30: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/30.jpg)
IPython is a more powerful interactive Python interpreter.
comments can include Latex math
[http://nbviewer.ipython.org/urls/raw.github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/master/Chapter1_Introduction/Chapter1_Introduction.ipynb]
![Page 31: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/31.jpg)
IPython is a more powerful interactive Python interpreter.
any plots generated by your code are displayed inline
[http://nbviewer.ipython.org/urls/raw.github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/master/Chapter1_Introduction/Chapter1_Introduction.ipynb]
![Page 32: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/32.jpg)
Write Python code interactively in a web browser instead of a terminal window.
![Page 33: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/33.jpg)
IPython has a decoupled client-server architecture.
[http://communities.intel.com/community/datastack/blog/2011/05/02/top-10-reasons-to-setup-a-client-server-network]
![Page 34: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/34.jpg)
server executes Python code
IPython has a decoupled client-server architecture.
[http://communities.intel.com/community/datastack/blog/2011/05/02/top-10-reasons-to-setup-a-client-server-network]
![Page 35: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/35.jpg)
clients see results of computation in real-time
IPython has a decoupled client-server architecture.
[http://communities.intel.com/community/datastack/blog/2011/05/02/top-10-reasons-to-setup-a-client-server-network]
![Page 36: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/36.jpg)
You can stay productive on any computer with an internet connection and a web browser.
[image source unknown]
![Page 37: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/37.jpg)
IPython integrates seamlessly with Matplotlib, making it well-suited for data exploration.
[http://matplotlib.org/gallery.html]
![Page 38: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/38.jpg)
Examples
[image source unknown]
![Page 39: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/39.jpg)
Resources
• http://nbviewer.ipython.org/4542975
• https://github.com/mroberts3000/IPythonIsGreat
• http://nbviewer.ipython.org/
• https://github.com/jrjohansson/scientific-python-lectures
• http://scipy-lectures.github.io/index.html
• Many more online and on the course web site
![Page 40: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/40.jpg)
Good Practices
• IPython notebooks to document your process
• Visualizations for data exploration
• Comment your code!
• Modularity - breaking down code into small functional, composable pieces
• Array-oriented computing
• Using assert statements and tests
• Version control (svn, git, github)
![Page 41: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/41.jpg)
Data Types
![Page 42: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/42.jpg)
Ben Shneiderman, 1996
• 1D (sequences)
• Temporal
• 2D (maps)
• 3D (shaped)
• nD (relational)
• Trees (hierarchical)
• Networks (graphs)
• Others?
The Eyes Have It: A Task by Data Type Taxonomy for Information Visualization [Shneiderman, 96]
![Page 43: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/43.jpg)
43
Tamara Munzner, 2013
![Page 44: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/44.jpg)
Semantics vs. Types
• Data Semantics: The real-world meaninge.g., company name, day of the month, person height, etc.
• Data Type: Interpretation in terms of scales of measurements
e.g., quantity or category, sensible mathematical operations, data structure, etc.
![Page 45: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/45.jpg)
![Page 46: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/46.jpg)
NominalCategoricalQualitative
Ordinal
Interval
Ratio
On the theory of scales and measurements [S. Stevens, 46]
![Page 47: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/47.jpg)
Data Types
• Nominal (Categorical) (N)Are = or ≠ to other values
Apples, Oranges, Bananas,...
• Ordinal (O)Obey a < relationship
Small, medium, large
• Quantitative (Q)Can do arithmetic on them
10 inches, 23 inches, etc.
On the theory of scales and measurements [S. Stevens, 46]
![Page 48: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/48.jpg)
Data Types• Q - Interval (location of zero arbitrary)
Dates: Jan 19; Location: (Lat, Long)
Like a geometric point. Cannot compare directly.
Only differences (i.e., intervals) can be compared
• Q - Ratio (zero fixed)Measurements: Length, Mass, Temp, ...
Origin is meaningful, can measure ratios & proportions
Like a geometric vector, origin is meaningful
On the theory of scales and measurements [S. Stevens, 46]
![Page 49: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/49.jpg)
Data Types
• N - Nominal (labels)Operations: =, ≠
• O - Ordinal (ordered)Operations: =, ≠, >, <
• Q - Interval (location of zero arbitrary)Operations: =, ≠, >, <, +, − (distance)
• Q - Ratio (zero fixed)Operations: =, ≠, >, <, +, −,×, ÷ (proportions)
On the theory of scales and measurements [S. Stevens, 46]
![Page 50: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/50.jpg)
Semantics
![Page 51: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/51.jpg)
Item
![Page 52: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/52.jpg)
Attributeaka
Feature
![Page 53: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/53.jpg)
1 = Quantitative2 = Nominal3 = Ordinal
![Page 54: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/54.jpg)
1 = Quantitative2 = Nominal3 = Ordinal
![Page 55: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/55.jpg)
Nominal /Ordinal = DimensionsDescribe the data, independent variables Quantitative = MeasuresNumbers to be analyzed, dependent variables
![Page 56: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/56.jpg)
Data vs. Conceptual Model
• Data Model: Low-level description of the data Set with operations, e.g., floats with +, -, /, *
• Conceptual Model: Mental constructionIncludes semantics, supports reasoning
Data Conceptual
1D floats temperature
3D vector of floats space
![Page 57: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/57.jpg)
Data vs. Conceptual Model• From data model...
32.5, 54.0, -17.3, … (floats)
• using conceptual model...Temperature
• to data typeContinuous to 4 significant figures (Q)
Hot, warm, cold (O)
Burned vs. Not burned (N)
Based on slide from Munzner
![Page 58: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/58.jpg)
Data Dimensions
![Page 59: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/59.jpg)
Univariate Data
Based on slide from M. Agrawala
http://www.smartmoney.com/marketmap/
![Page 60: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/60.jpg)
Bivariate Data
Based on slide from M. Agrawala
Scatterplot is common
![Page 61: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/61.jpg)
Trivariate Data
Based on slide from M. Agrawala
Do NOT use 3D scatterplots!
![Page 62: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/62.jpg)
Trivariate Data
Based on slide from M. Agrawala
Map the third dimension to some other visual attribute
![Page 63: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/63.jpg)
Multivariate Data
Based on slide from M. Agrawala
Give each attribute its own display (small multiples)
![Page 64: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/64.jpg)
Multivariate Data Representations
![Page 65: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/65.jpg)
Data Reduction
• Filtering: Eliminate some items or attributese.g., select range of interest, zoom in, remove outliers, etc.
• Aggregation: Represent a group of elements by a new derived element
e.g., take average, min, max, count, sum
Attribute aggregation a.k.a. dimensionality reduction
![Page 66: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/66.jpg)
Mapping Data to Images
![Page 67: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/67.jpg)
13
Image
Visual language is a sign system
Images perceived as a set of signs
Sender encodes information in signs
Receiver decodes information from signs
Semiology of Graphics, 1983
Jacques Bertin
Jacques Bertin
• French cartographer [1918-2010]
• Semiology of Graphics [1967]
• Theoretical principles for visual encodings
![Page 68: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/68.jpg)
Bertin’s Visual AttributesPoints Lines AreasMarks
Position
Size
(Grey)Value
Texture
Color
Orientation
Shape
Channels
Bertin, Semiology of Graphics, 1967
![Page 69: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/69.jpg)
Large Design Space (Visual Metaphors)
J. Bertin, “Graphics and Graphic Information Processing”, 1981
![Page 70: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/70.jpg)
Example
W. Playfair, 1786
![Page 71: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/71.jpg)
Example
W. Playfair, 1786
x-axis: Year (Q)y-axis: Currency (Q)Color: Imports / Exports (N, O)
![Page 72: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/72.jpg)
Effective Visual Attributes
![Page 73: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/73.jpg)
Compare These Values
Cole Nussbaumer
![Page 74: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/74.jpg)
Order These Colors
Based on slide from Stasko
![Page 75: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/75.jpg)
Order These Colors
Based on slide from Stasko
![Page 76: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/76.jpg)
Order These Colors
Based on slide from Stasko
![Page 77: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/77.jpg)
Brightness Saturation Hue: not as muchPerceived as Ordered
![Page 78: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/78.jpg)
Visual Attributes per Data Type
Jock Mackinlay “Automating The Design of Graphical Presentations.” 1986
Bertin, 1967
William S. Cleveland; Robert McGill , “Graphical Perception: Theory,
Experimentation, and Application to the Development of Graphical Methods.” 1984
Cleveland / McGill, 1984 Mackinlay, 1986
Bertin, Semiology of Graphics, 1967
![Page 79: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/79.jpg)
Most Efficient
Least Efficient
C. MulbrandonVisualizingEconomics.com
Quantitative
Ordinal
Nominal
}}}
![Page 80: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/80.jpg)
Most Effective
VisualizingEconomics.com
![Page 81: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/81.jpg)
Less Effective
VisualizingEconomics.com
![Page 83: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/83.jpg)
Effective Visualizations
![Page 84: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/84.jpg)
Not Effective...
Sources: US Treasury and WHO reports
![Page 85: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/85.jpg)
Also not effective...
Source: Nature
![Page 86: 02 Process](https://reader033.fdocuments.us/reader033/viewer/2022042615/563dbb99550346aa9aae99bf/html5/thumbnails/86.jpg)
Much better...
Sources: US Treasury, WHO, Nature