-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathschool.html
166 lines (138 loc) · 8.14 KB
/
school.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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Alex's Portfolio | Past Projects</title>
<link rel="icon" href="assets/notebook.svg">
<link rel="stylesheet" href="style.css">
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-B9FXS221SP"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-B9FXS221SP');
</script>
</head>
<body>
<div>
<h1 id="title">Alex's Portfolio</h1>
</div>
<div id="nav">
<h2>
<a href="index.html" class="nav">About</a>
<a href="resume.html" class="nav">Resume</a>
<a href="projects.html" class="nav">Projects</a>
<a href="school.html" class="nav" id="active">School Projects</a>
</h2>
</div>
<div class="body">
<h2 id="visualization" class="sec"><strong>Visualization</strong></h2>
<div class="card">
<h3 id="visualization-and-exploration-of-the-gapminder-dataset-using-ggplottidyverse-packages-in-r"
class="proj">
<em>Visualization and Exploration of the Gapminder Dataset Using ggplot/tidyverse Packages in R</em>
</h3>
<hr>
<p>View this project’s <a
href="https://github.com/cluffa/stat5730project/blob/master/final_report_Alex_Cluff.pdf">final
report</a> and <a
href="https://github.com/cluffa/stat5730project/blob/master/final_report_Alex_Cluff.Rmd">source
code</a>.</p>
<p><img src="https://github.com/cluffa/stat5730project/raw/master/final_report_Alex_Cluff_files/figure-gfm/unnamed-chunk-4-1.png"
alt="ggplot graph" /></p>
<p>The focus of this project was to explore the famous Gapminder dataset visually using ggplot graphs like
the one above. I also used the other tidyverse packages like dplyr to manipulate the data in different
ways
to create well formatted data that fits into summary tables.</p>
</div>
<h2 id="time-series-analysis" class="sec"><strong>Time Series Analysis</strong></h2>
<div class="card">
<h3 id="analysis-of-homicides-in-the-us-over-time-using-r-and-an-armasarima-model" class="proj"><em>Analysis
of
Homicides in
the US Over Time Using R and an ARMA/SARIMA Model</em></h3>
<hr>
<p>View this project’s <a href="https://cluffa.github.io/stat5550project/">final report</a> and <a
href="https://github.com/cluffa/stat5550project/blob/master/final_project_Alex_Cluff.Rmd">source
code.</a><br><br>
<img src="assets/ts_plot.png" alt="homicide predictions" />
</p>
<p>This project was based around forecasting methods. I find overall trends, monthly seasonality, and fit
ARIMA and SARIMA models. I compare the two model’s performance as well as fit. I then forecasted
homicides for
the next 24 months.</p>
</div>
<h2 id="regression-analysis" class="sec"><strong>Regression Analysis</strong></h2>
<div class="card">
<h3 class="proj"
id="reproducing-the-results-and-logistic-regression-model-of-a-study-on-modeling-prison-sentencing-from-facial-features">
<em>Reproducing the Results and Logistic Regression Model of a study on Modeling Prison Sentencing From
Facial Features</em>
</h3>
<hr>
<p>View this project’s <a href="https://github.com/cluffa/stat3302project/blob/main/group_project.pdf">final
report</a> and <a href="https://github.com/cluffa/stat3302project/blob/main/model.R">source
code</a>.
</p>
<p><img src="https://github.com/cluffa/stat3302project/blob/main/table2.png?raw=true"
alt="model coefficient table" /></p>
<p>The idea for this assignment was to gain experience and become more comfortable reading and interpreting
scientific research papers. We also learned the importance of reproducibility and transparency. My group
was tasked with reproducing the results and models from this paper and reporting on them. I was in
charge of
the modeling as well as the table for the models, both of which are created with the source code I
linked.
The picture above is a replication of the table used in the original paper. Interestingly, we ended up
finding a small mistake in the paper.</p>
</div>
<h2 id="machine-learning-1" class="sec"><strong>Machine Learning</strong></h2>
<div class="card">
<h3 id="predicting-ebay-car-prices-using-a-random-forest-model-in-r" class="proj"><em>Predicting Ebay Car
Prices Using a Random Forest Model in R</em></h3>
<hr>
<p>View this project’s <a
href="https://github.com/cluffa/stat4620project/blob/master/final_report_made_in_colaboration_with_classmates.pdf">final
report</a>, <a href="https://github.com/cluffa/stat4620project/blob/master/randomForest.R">source
code
for the random forest model</a>, and <a
href="https://github.com/cluffa/stat4620project/blob/master/clean_autos_dataset.R">source code for
cleaning the data</a>.</p>
<p>This was a group project. I handled the random forest model as well as the data cleaning. We each tried a
model and compared results. The random forest model came out on top based on testing MSE. I had to learn
a lot about resource allocation to complete this project. The dataset had 180k rows and 10+ possible
predictors. I quickly found out that I would not be able to easily tune and train the model. The
training process ended up taking a few days with cross-validation on less than half of the full dataset.
I had to
weigh run time vs accuracy and pick model parameters early on in the training process.</p>
</div>
<div class="card">
<h3 class="proj"
id="fitting-a-bayesian-hierarchical-model-on-fake-flu-data-simulated-with-an-mcmc-algorithm-using-r-and-rjagsjags">
<em>Fitting a Bayesian Hierarchical Model on Fake Flu Data. Simulated with an MCMC algorithm using R and
Rjags/jags.</em>
</h3>
<hr>
<p>View this project’s <a
href="https://github.com/cluffa/stat3303project/blob/master/Final_project_Alex_Cluff.pdf">final
report</a>, <a
href="https://github.com/cluffa/stat3303project/blob/master/Final_project_Alex_Cluff.Rmd">report
source
code</a>, and <a href="https://github.com/cluffa/stat3303project/blob/master/fit.R">model fitting
source
code</a></p>
<p>The setup for this project:<br />
“There are two tests for influenza strain K9C9. The data collected consists of 10 countries and 100
pairs of test results. The more accurate of the tests will be assumed fact. The less accurate test, EZK,
is the
area of interest for this project. A Bayesian hierarchical model will be fit and it will be simulated
with an
MCMC algorithm using R/jags.”<br />
I fit the model, assessed fit, and interpreted the results in the context of a global pandemic.</p>
</div>
</div>
<script src="script.js"></script>
</body>
</html>