-
Notifications
You must be signed in to change notification settings - Fork 0
/
covid.html
237 lines (211 loc) · 16.2 KB
/
covid.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
<!DOCTYPE html>
<html lang="en">
<head>
<link rel="stylesheet" type="text/css" href="vendors/css/normalize.css">
<link rel="stylesheet" type="text/css" href="resources/css/styles-projects.css">
<link rel="stylesheet" type="text/css" href="vendors/css/ionicons.min.css">
<link rel="stylesheet" type="text/css" href="vendors/css/grid.css">
<link rel="preconnect" href="https://fonts.gstatic.com">
<link href="https://fonts.googleapis.com/css2?family=Montserrat:wght@300&display=swap" rel="stylesheet">
<link rel="apple-touch-icon" sizes="180x180" href="vendors/logos/apple-touch-icon.png">
<link rel="icon" type="image/png" sizes="32x32" href="vendors/logos/favicon-32x32.png">
<link rel="icon" type="image/png" sizes="16x16" href="vendors/logos/favicon-16x16.png">
<link rel="manifest" href="vendors/logos/site.webmanifest">
<link rel="mask-icon" href="vendors/logos/safari-pinned-tab.svg" color="#006266">
<meta name="msapplication-TileColor" content="#006266">
<meta name="theme-color" content="#ffffff">
<title>Data Analytics and Visualization | Sophie Sun | Projects</title>
</head>
<body>
<header>
<nav class="sticky">
<div class="row">
<a href="index.html"><img src="vendors/logos/android-chrome-512x512.png" alt="sophielogo" class="logo"></a>
<ul class="top-nav">
<li><a href="index.html">Home</a></li>
<li><a href="projects.html">Projects</a></li>
<li><a href="visuals.html">Visuals</a></li>
<li><a href="skills.html">Skills</a></li>
<li><a href="bio.html">Bio</a></li>
</ul>
</div>
</nav>
</header>
<!---sidebar -->
<div id="menu">
<div class="nav vertical-1-of-2">
<div class="consumerstudies">
<h3>Consumer Studies</h3>
<div class="sidebar">
<a href="mask.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/mask.jpg" alt="mask_design_intelligence" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Mask Design Intel</h4>
<p>Optimize product with machine learning from Amazon reviews</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="projects.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/cookie.jpg" alt="finding_perfect_chocolate_chip_cookie" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Winner Cookie Recipe</h4>
<p>Find the perfect chocolate cookie recipe through experiments</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="shirtsize.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/shirtsize.jpg" alt="finding_shirt_size_for_CA_consumers" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Shirt Size Ratio</h4>
<p>Quantify the shirt size ratio for CA consumers based on surveys</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="scentedtextile.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/textile.jpg" alt="scented_fabric_preference" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Scented Textile</h4>
<p>Discover the cultural influence on scented textile preference</p>
</div>
</a>
</div>
<p> ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅</p>
</div>
</div>
<div class="nav vertical-1-of-2">
<div class="socialstudies">
<h3>Social Studies</h3>
<div class="sidebar">
<a href="romance.html" class="project-photo">
<div class="col span-1-of-4">
<img src="projects/7romance/lovesongtopwords.png" alt="love song lyrics analysis" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Romance Thermometer</h4>
<p>Analyze lyrics over nine decades</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="covid.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/covid.jpg" alt="evaluating_covid_vaccine_acceptance_factors" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Vaccine Acceptance</h4>
<p>Find out what encourage people to get vaccinated</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="pandemic.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/pandemic.jpg" alt="covid-19 pandemic" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>COVID-19 Pandemic</h4>
<p>First end-of-year report on the global fight with covid</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="hemp.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/hemp.jpg" alt="hemp" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Drug Consumption</h4>
<p>Sketch out demo|psycho-graphic traits of marijuana users</p>
</div>
</a>
</div>
<div class="sidebar">
<a href="mathscore.html" class="project-photo">
<div class="col span-1-of-4">
<img src="vendors/images/school.jpg" alt="school_performance" style="width:100%">
</div>
<div class="col span-3-of-4">
<h4>Math Score</h4>
<p>Evaluate factors on high schooler's math performance</p>
</div>
</a>
</div>
<p> ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅</p>
</div>
</div>
</div>
<!---main body -->
<section class="project-demo">
<div class ="col span-1-of-4">
<div class="highlights">
<h5>Year</h5>
<p>2020</p>
<h5>Scope</h5>
<p>5 sets of questionnaires<br>on 7-point Likert scale<br>surveyed in summer 2020</p>
<h5>Variables</h5>
<p>75 variables<br>vaccine attributes<br>demographic data<br>political data</p>
<h5>Observations</h5>
<p>1971</p>
<h5>Script</h5>
<p>R<br>sensemakr</p>
<h5>Method</h5>
<p>Matching with PS and IPW<br>Sensitivity Analysis</p>
</div>
</div>
<div class ="col span-3-of-4">
<div class="project-details">
<h2>What Makes People Willing To Vaccinate?</h2>
<p>To make causal inference from observational data, we worked under Selection On Observables (SOO) with two assumptions: conditional ignobility and common support. In other words, selecting a subset of observations where all else being equal statistically, we can then study the treatment factor being randomly present and measure the difference in effect between the treated and untreated. It all comes down to the fundamental rule of making a causal inference: treated effect Y1 - untreated effect Y0 conditioned on comparable X. </p>
<img src="projects/4covid/SOO.png" style="width:100%">
<p>Individuals have many attributes - demographic (age, gender, race, education, income) or ideologic (religious beliefs, political leaning, trust in big Pharma, approval of federal government being Trump administration at the time of the survey) or situational (if they know people who get severely sick or dead from covid, or if they can work from home) - and each attribute can have many discrete levels or continuous values. The statistical principle used in this project is Matching. Match all other attributes and leave only the one as treatment factor before measuring the difference that one treatment factor makes. Then that one factor was gauged with other known strong factors to quantify the power of the causal effect.</p>
<h5>Findings</h5>
<p>Several factors are found to be significant in affecting individuals’ choice on vaccine acceptance. I’ll show only one treatment factor studied here: knowing someone in your life hospitalized or dead from covid-19. We can see how our personal experience of how fatal or damaging the virus can cause, not reported on the news, but on someone we know, can have a big push over our vaccine decision. But how much of that force? And can we prove it’s really a push? </p>
<h5>Design</h5>
<p>In the practice of Selection-on-Observables identification strategy, we applied a few Matching methods: a. propensity scores calculated through logistic regressions on other covariates, and b. weighting with inverse propensity weights (IPW). So we have an “apples to apples” comparison: people with very similar attributes computed by the propensity score in a similar fashion to principle component analysis (PCA) which collapse multi dimensions into one value, and then we check on the difference of whether or not knowing someone sick or dead from covid can make them score 1-7 on willingness for vaccination. </p>
<p>Before matching, the propensity scores between the treated and untreated look like this:</p>
<img src="projects/4covid/knowcovidsevere_ps.png" style="width:100%">
<p>Here we can see the magic of Matching: blue dots are BEFORE matching mean difference between treated and untreated on covariates. It’s all over the place. Orange dots are AFTER matching. The mean difference are much smaller. We found similar apples to apples comparison on that one factor.</p>
<img src="projects/4covid/knowcovidsevere_balance.png" style="width:100%">
<p>After we compared apple to apple, we can see from statistical measurements that given all other covariates, knowing someone with severe health results form Covid has a t-value at 5.384 and for a confounder to explain away what was explained by the treatment at 95% confidence interval, that confounder has to make up 2.4% of residual variance for both the treatment and the outcome.</p>
<p>Another visual way to check how effective this treatment on the outcome is by Sensitive Analysis plotted below. Here we use known factors such as attitude toward mandatory vaccination as benchmark to compare the knowing someone severely sick or dead from the covid factor with. The red dotted line is when the effect of the treatment goes to zero. It is measured as 8 times the effect of mandatory vaccination or 5 times of the trust of pharmaceutical companies. </p>
<img src="projects/4covid/sensitivity.png" style="width:100%">
<p>With all the above quantitative analysis, we can conclude that knowing someone severely sick or dead from covid can make people more willing to take vaccines.</p>
</div>
</div>
</section>
<footer class="footer">
<div class="row">
<div class ="col span-1-of-2">
<ul class="footer-nav">
<li><a href="index.html" style="text-decoration: none">Home</a></li>
<li><a href="projects.html" style="text-decoration: none">Projects</a></li>
<li><a href="visuals.html" style="text-decoration: none">Visuals</a></li>
<li><a href="skills.html" style="text-decoration: none">Skills</a></li>
<li><a href="bio.html" style="text-decoration: none">Bio</a></li>
</ul>
</div>
<div class ="col span-1-of-2">
<ul class="socialmedia">
<li><a href="https://github.com/sophiexsun" target="_blank" rel="noopener noreferrer"><img src="vendors/icons/logo-github.svg", alt="github icon", class="link-icon"></a></li>
<li><a href="https://www.linkedin.com/in/sophiexsun/" target="_blank" rel="noopener noreferrer"><img src="vendors/icons/logo-linkedin.svg", alt="linkedin icon", class="link-icon"></a></li>
<li><a href="https://www.youtube.com/channel/UCdviDsjKArhG5GXRIPzJtrw" target="_blank" rel="noopener noreferrer"><img src="vendors/icons/logo-youtube.svg", alt="youtube icon", class="link-icon"></a></li>
</ul>
</div>
</div>
<div class="row">
<p> © 2021-2022 by Sophie X Sun. All Rights Reserved. </p>
</div>
</footer>
</body>
</html>