EngineeringProject/index.html at master · arunsingh100/EngineeringProject · GitHub

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
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
<!DOCTYPE html>
<html lang="en">
<head>
    <title>Chlorophyll + Sea Surface Temperatures</title>

   <!-- Javascript -->
    <script src="https://code.jquery.com/jquery-2.2.2.min.js" integrity="sha256-36cp2Co+/62rEAAYHLmRCPIych47CvdM+uTBJwSzWjI=" crossorigin="anonymous"></script>
    <script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.6/js/bootstrap.min.js" integrity="sha384-0mSbJDEHialfmuBBQP6A4Qrprq5OVfW37PRR3j5ELqxss1yVqOtnepnHVP9aJ7xS" crossorigin="anonymous"></script>
    <script src='https://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML'></script><!-- MathJAX -->
    <script src="js/parallax.js"></script><!-- Parallax Banner -->
    <script src="js/navbar.hide.js"></script><!-- Hide Navbar -->
    <script src="js/scroll.js"></script><!-- Affix Sidebar/Scroll Functions -->

    <!-- CSS -->
    <link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.6/css/bootstrap.min.css" integrity="sha384-1q8mTJOASx8j1Au+a5WDVnPi2lkFfwwEAa8hDDdjZlpLegxhjVME1fgjWPGmkzs7" crossorigin="anonymous">
    <link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/font-awesome/4.5.0/css/font-awesome.min.css">
    <link href='https://fonts.googleapis.com/css?family=Coda:400,800' rel='stylesheet' type='text/css'><!-- Google Font for Title -->
    <link rel="stylesheet" type="text/css" href="css/cdup_tutorial.css"><!-- Custom Theme for know.data tutorial -->

    <!-- Syntax Highlighting -->
    <!-- Support for the following languages: -->
    <!-- Apache, Bash, C#, C++, CSS, CoffeeScript, Device Tree, Diff, HTML, XML, HTTP, Ini, JSON, Java, JavaScript, Makefile, Markdown, Nginx, Objective-C, PHP, Perl, Python, Ruby, SQL, Fortran, Julia, Lisp, Lua, Mathematica, Matlab, Python-Profile, R, Scilab, Scala, Stata, Swift -->
    <link rel="stylesheet" type="text/css" href="css/styles/github.css"><!-- Style for highlighting Code: Default to Github -->
    <script src="js/highlight.pack.js"></script>
    <script>hljs.initHighlightingOnLoad();</script><!-- Activate Code Highlighting -->

</head>
<body>
    <nav id='navbar' class="nav navbar-default navbar-fixed-top navbar-border"><!-- Navbar -->
        <div class="container-fluid">
            <div class="navbar-header">
                <button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#collapse-links" aria-expanded="false">
                    <span class="sr-only">Toggle navigation</span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                </button>
                <a class="navbar-brand navbar-color" href="https://commerce.gov/datausability"><strong>Data Usability</strong></a>
            </div> <!-- navbar-header -->

            <!-- Collect the nav links, forms, and other content for toggling -->
            <div class="collapse navbar-collapse" id="collapse-links">
                <ul class="nav navbar-nav navbar-color">
                    <li class="disclaimer"><a href="https://commerce.gov/datausability">a project by Commerce Data Service</a></li>
                </ul>
                <ul class="nav navbar-nav navbar-right navbar-color">
                    <li><a href="https://www.commerce.gov/datausability/">Index</a></li>
                </ul>
            </div><!-- navbar-collapse -->
        </div><!-- container-fluid -->
    </nav>

    <!-- Banner -->
    <section class="scroll">
        <div class="scroll-overlay ">
            <div class="title headtext">
                <h1><span class="title-line" style="font-size:150%;">Predicting Markers for Environmental Impact</span></h1>
                <h4><span class="title-line">Modelling Relationships Between Chlorophyll and Sea Surface Temperatures</span></h4>
                <h5><span class="title-line">by Nikash Sethi and Arun Singh (Department of Commerce) </span></h5>
            </div>
        </div>
    </section>

    <!-- Body -->

    <section>
        <div class="container-fluid content">
            <div class="row">


                <div id='' class="hidden-xs hidden-sm col-lg-4 col-md-4">
                    <a href="https://commerce.gov/dataservice"><img class="footlogo" width="160px"src="img/CDS-horizontal-v2.jpg"/></a>

                </div>
                <div id='content' class="col-lg-6 col-md-6">
                    <em>  As part of the <strong><a href="https://www.commerce.gov/datausability/">Commerce Data Usability Project</a></strong>, the <a href="https://www.commerce.gov/dataservice/">Commerce Data Service</a> held a Summer Internship for high school students in 2016. This is the culmination of a research project conducted by 2 summer associates of that program under Chief Data Scientist Jeff Chen. If you have questions, feel free to reach out to the Commerce Data Service at <a href="mailto:DataUsability@doc.gov">DataUsability@doc.gov</a></em>
                </div>
            </div>
            <hr>

            <br>

            <div class="row">
                <div id='content' class="col-lg-10 col-md-10"><!-- Content -->

                <section id='intro'>
                    <h2 class="sectionhead">Atmospheric Data is a Digital Representation of the Living Environment.</h2>


                <p>  In 2016, NASA released a 3-D visualization that captured the patterns of carbon dioxide in the atmosphere. Through various ground sensors, scientists showed the rising concentrations of carbon dioxide throughout the globe. Because of an intense increase in the burning of fossil fuels for energy, greenhouse gases have also risen in concentration, warming the Earth's atmosphere day by day. Scientists believe that a quarter of these emissions are being absorbed by the ocean, and have the potential to kill microorganisms that are the basis of aquatic and land ecosystems. One essential organism in danger is phytoplankton. As a primary producer in the ocean, phytoplankton are responsible for cycling carbon dioxide to regions from the atmosphere to the depths of the ocean for other organisms to feed off and reproduce (Lindsey, 2010). These creatures also heavily impact the carbon cycle and surface temperatures.</p>

                <iframe width="950" height="515" src="https://www.youtube.com/embed/syU1rRCp7E8?ecver=1" frameborder="0" allowfullscreen></iframe>

                <p>We were curious to see if any relationship existed between these two inversely related events. In reviewing the research, we have found that a) temperature levels have risen over the past 30 years, and b) phytoplankton levels have decreased over the last 40 years (Biello, 2010). Our goal was to correlate this data to suggest that an effort be made towards increasing algal phytoplankton blooms in oceans for a decrease in carbon dioxide, leading to cooler temperatures for the atmosphere. Ultimately, after performing regressions on the data, we could determine that it is possible to predict Chlorophyll levels based on temperature to a certain degree of accuracy.</p>

                <br><img src="img/carboncycle.png" alt="screencap"/><br><br>
                <br><img src="img/outsidedata.png" alt="screencap"/><br><br>


                </section>
               <section id='data'>
                    <h2 class="sectionhead">Data as a public good</h2>
                <p>Our tutorial was developed mostly through HTML and JavaScript. CSS design files were integrated with the application that we developed to maintain the style of the websites that CDS publishes. Ultimately, we developed simpler graphs within our Python program using MatPlotLib, and included images of those graphs in our tutorial.
                </p>

                <p>Next, we selected specific regions of data that we were going to analyze, each region consisting of 16 coordinate points in the shape of either a square or a rectangle (ultimately, we would like to analyze every data point in the ocean using GIS Shapefiles to select geographic regions, but since time was a limitation, we chose not to do so).

                We chose three locations in which the levels of phytoplankton would vary enough to provide a generalization of the trends that occur in oceans. </p>

                <ul>
                    <li>Since phytoplankton levels decrease farther away from the coast, our first region was in the middle of the Atlantic Ocean, where the Chlorophyll levels were almost 0.</li>
                    <li>Next, we chose to analyze a region in the Mediterranean Sea, as it had medium levels of Chlorophyll. </li>
                    <li>Finally, our last region was near the coast of Argentina, where an abundance of Chlorophyll was present in the ocean.
                    </li>
                </ul>

                <p>
                  We determined these regions based on latitude and longitude coordinates, by outlining the areas on a map and converting the coordinates into sets of indices that could select certain data from the overall datasets.
                </p>

                <p>
                After we chose the points that we wanted to analyze, we obtained the Chlorophyll and temperature values for each point from our datasets. We then stored each of these time series in Numpy arrays, which are structures that are useful in accessing large lengths of data. After doing so, we plotted the Chlorophyll levels against the temperature values, to accurately represent and visualize the relationships between the two variables. These graphs, which were coded using the MatPlotLib PyPlot library in Python, are shown below:
                </p>


               </section>
                <section id='code'>
                    <h2 class="sectionhead">Getting Started</h2>

                <p>The data that we used was downloaded from an online NASA satellite database: we selected Chlorophyll and sea surface temperature datasets for each month from 2003 to 2015. Each dataset was in the form of netCDF. This datatype was not intuitive to use, and so we employed the netCDF4 Python module to be able to read and access data. Once we stored the datasets locally, we referenced various netCDF and specific NASA APIs to familiarize ourselves with the format and structure of the data.
                </p>
                <p>To get started with the data, this tutorial illustrates how to replicate the analysis done for the three regions of Chlorophyll vs. Sea Surface Temperature. For reference to the data from NASA, please visit the <a href="https://neo.sci.gsfc.nasa.gov/view.php?datasetId=MY1DMM_CHLORA">Chlorophyll</a> or <a href="https://neo.sci.gsfc.nasa.gov/view.php?datasetId=MYD28M"> Sea Surface Temperature Data</a>.</p>


                </section>

                <section id="visual">
                    <h2 class="sectionhead">Part 1: Preliminaries</h2>
                    <p>First, you need to import the following libraries:</p>

                    <ul>
                        <li><strong>netCDF4</strong>. A Python interface to the netCDF library.</li>
                        <li><strong>numpy</strong>. An extremely powerful data storage package that is beneficial when working with large, local datasets. .</li>
                        <li><strong>matplotlib</strong>. Helpful to create visualizations and outline trends in datasets.</li>
                        <li><strong>scipy</strong>. A library that contains modules for technical and mathematical computing. </li>

                    </ul>
<pre class="r"><code>#from scipy.io.netcdf import netcdf_file as Dataset
<b>from</b> netCDF4 <b>import</b> Dataset
<b>import</b> numpy <b>as</b> np
<b>import</b> numpy.fft <b>as</b> fft
<b>import</b> os
<b>import</b> matplotlib.pyplot <b>as</b> plt
<b>import</b> math
<b>from</b> scipy <b>import</b> stats
<b>from</b> scipy.optimize <b>import</b> curve_fit
<b>from</b> scipy.optimize <b>import</b> leastsq
<b>from</b> scipy.signal <b>import</b> correlate

#from scipy import optimize
<b>import</b> json</code></pre>

                    </section>

                <section id="conclusion">
                    <h2 class="sectionhead">Part 2: Understanding the Data</h2>

                <p> According to NEO NASA, the Sea Surface Temperature Data provides temperature levels at 1km, 4.6 km, and 36 km resolutions over the Earth's oceans. Measurements from the Moderate Resolution Imaging Spectroradiometer (MODIS) instruments measure the "warmth of the ocean's "skin" (top millimeter)". </p>

                <br><img src="img/sst.png" alt="screencap"/><br><br>


                <p>The Chlorophyll Data from NEO NASA is measured at a 1 km resolution, restricted to clear water with daily coverage via the MODIS instruments on NASA's Terra and Aqua satellites. </p>

                <br><img src="img/chloro.png"/><br><br>


                <p>To learn more about this dataset in depth, please visit NEO NASA's Earth Observations for <a href="https://neo.sci.gsfc.nasa.gov/view.php?datasetId=MY1DMM_CHLORA">Chlorophyll</a> or <a href="https://neo.sci.gsfc.nasa.gov/view.php?datasetId=MYD28M"> Sea Surface Temperature Data</a>.</p>


                <h3><strong>Part 2.1: Methods</strong></h3>
                <p>These are the methods to get the Latitude, Longitude, Chlorophyll and Temperature data.</p>

<pre class="r"><code>def getLatLong(lats, lons, actualLat, actualLon): #determine lat/long given indeces in array
    lonPlace = float("inf")
    lonsCount = 0
    for thing in lons:
        if (math.fabs(thing-actualLon) < 0.1):
            lonPlace = lonsCount
            break
        lonsCount += 1
    latPlace = float("inf")
    latsCount = 0
    for thing in lats:
        if (math.fabs(thing-actualLat) < 0.1):
            latPlace = latsCount
            break
        latsCount += 1
    return (lonPlace, latPlace)</code></pre>

<pre class="r"><code>def getChlorophyllData(latitude, longitude, chloroFiles): #store chlorophyll data for given location in array
    chloroList = []
    for chloras in chloroFiles:
        value = chloras[longitude][latitude]
        chloroList.append(value)
    return (chloroList)</code></pre>


<pre class="r"><code>def getTemperatureData(latitude, longitude, tempFiles): #store temperature data for given location in array
    tempList = []
    for temps in tempFiles:
        value = temps[longitude][latitude]
        tempList.append(value)
    return (tempList)</code></pre>


<pre class="r"><code>def getDataFromRange(latOne, latTwo, lonOne, lonTwo, chloroFiles, tempFiles, lons, lats):
    allChloroList = []
    allTempList = []
    for lat in range(latOne, latTwo):
        for lon in range(lonOne,lonTwo):
            (mylatitude, mylongitude) = getLatLong(lats,lons,lat,lon)
            chloroList = getChlorophyllData(mylatitude, mylongitude, chloroFiles)
            tempList = getTemperatureData(mylatitude, mylongitude, tempFiles)
            allChloroList.extend(chloroList)
            allTempList.extend(tempList)

    actualAllChloro = []
    actualAllTemp = []
    timeseries = []
    for i in range(len(allChloroList)):
        if not (allChloroList[i] is np.ma.masked or allTempList[i] is np.ma.masked):
            actualAllChloro.append(allChloroList[i])
            actualAllTemp.append(allTempList[i])
            timeseries.append(i)

    """allChloroList=[]
    allTempList=[]
    for i in range(len(actualAllTemp)):
        if not (actualAllTemp[i]is np.ma.masked):
            allTempList.append(actualAllTemp[i])
            allChloroList.append(actualAllChloro[i])"""

    return (actualAllChloro, actualAllTemp, timeseries)</code></pre>


<pre class="r"><code>def getAllFiles(): #initiate data files
    chloroFiles = []
    count = 0
    for chloroFile in os.listdir('chlorophyllData'):
        chloroFilename = "chlorophyllData/" + str(chloroFile)
        chloroData = Dataset(chloroFilename, mode="r")
        if count == 0:
            lons = chloroData.variables['lon'][:]
            lats = chloroData.variables['lat'][:]
        #lons = chloroData.variables['lon'][:]
        #lats = chloroData.variables['lat'][:]
        chloras = chloroData.variables['chlor_a'][:]
        chloroFiles.append(chloras)
        print (count)
        count += 1
    tempFiles = []
    count = 0
    for tempFile in os.listdir('temperatureData'):
        tempFilename = "temperatureData/" + str(tempFile)
        tempData = Dataset(tempFilename, mode="r")
        #lons = tempData.variables['lon'][:]
        #lats = tempData.variables['lat'][:]
        temps = tempData.variables['sst'][:]
        tempFiles.append(temps)
        print (count)
        count += 1
    return (chloroFiles, tempFiles, lons, lats)</code></pre>

                <br><img src="img/atlanticmap.png"/><br><br>
                    <p>Center of the Atlantic Ocean zone. </p>
                <br><img src="img/medmap.png"/><br><br>
                    <p>Center of the Mediterranean Sea zone. </p>
                <br><img src="img/argentinamap.png"/><br><br>
                    <p>Center of the Coastline of Argentina zone. </p>


                    <h3><strong>Part 2.2: Plot Data </strong></h3>
                <p>We had to find a way to uniquely store data and make a visualization to plot the points. The visualization, a simple graph created through cv2 module in Python, contains points for sea surface temperature and chlorophyll levels for each time period (every month since January 2003).This graph will serve two purposes: it allows us to make sure that the data was entered into the Numpy arrays correctly, and it will help decide what type of statistical analysis technique to use. </p>

                <p> To create the graph, we stored the chlorophyll and temperature values at that point for each month over three years in an array, and graphed the array using Pyplot. From the graph, we could see that both variables were sinusoidal because they vary depending on the season. Furthermore, they were inverses of each other: when temperatures were higher, chlorophyll levels were lower, and vice versa. However, the trends of chlorophyll had a much less definitive sinusoidal relationship, as they varied within their overall sinusoidal shapes.</p>


                <pre class="r"><code>#SET STYLE OF MATPLOTLIB GRAPH
                plt.style.use('ggplot')

                #PLOT OF CHLOROPHYLL VS. TIME (MIDDLE OF OCEAN)
                plt.plot(allChloroList1, linestyle='-', color='g')
                plt.xlabel("Time (Months Since January 2003)")
                plt.ylabel("Chlorophyll Levels (mg/m^3)")
                plt.suptitle("Chlorophyll Levels Monthly Over the Course of 12 Years")
                plt.figure()

                #PLOT OF TEMPERATURE VS. TIME (MIDDLE OF OCEAN)
                plt.plot(allTempList1, linestyle='-', color='r')
                plt.xlabel("Time (Months Since January 2003)")
                plt.ylabel("Sea Surface Temperature (Degrees Celcius)")
                plt.suptitle("Sea Surface Temperature Monthly Over the Course of 12 Years")
                plt.figure()

                </code></pre>

                <br><img src="img/atlanticdata.png"/><br><br>
                <br><img src="img/meddata.png"/><br><br>
                <br><img src="img/argentinadata.png"/><br><br>

                    <h3><strong>Part 2.3: Plot Regression over Data </strong></h3>

                <p>Because of this triangular trend, we predicted that although a linear regression would not accurately fit the scatter plot, the residuals from each point to that regression line would form a harmonic, oscillating pattern that would be neatly modeled through a sinusoidal regression. Therefore, we conducted the linear regression (Figure 9) on each of the 3 regions using the SciPy package in Python. Next, we iterated through each point and calculated its residual with the regression that we conducted. We stored these residuals in a time series plot, and as we predicted, they maintained a harmonic trend (Figure 10). After performing a harmonic regression on the residual plot, we had an accurate approximation of the error from the linear regression for each point. These regressions are outlined in the visualizations below, which demonstrate the linear regression on Chlorophyll and temperature data in the Atlantic Ocean, followed by a harmonic regression on the residuals of the data. </p>

<pre class="r"><code>#PERFORM LINEAR REGRESSION ON DATA
regression1 = np.polyfit(allTempList1, allChloroList1, deg=1)

#PLOT LINEAR REGRESSION
plt.plot(allTempList1, regression1[0]*allTempList1+regression1[1], linewidth='5')

#PLOTTED ON TOP OF SCATTER OF TEMPERATURE VS. CHLOROPHYLL
plt.scatter(allTempList1, allChloroList1, color='c')
plt.xlabel("Sea Surface Temperature (Degrees Celcius)")
plt.ylabel("Chlorophyll Levels (mg/m^3)")
plt.suptitle("Temperature (Middle of Atlantic Ocean) vs. Chlorophyll")
plt.figure()
                </code></pre>

<pre class="r"><code>#CALCULATE RESIDUALS (ERRORS OF LINEAR REGRESSION)
residualsArray1 = []
for i in range(len(allTempList1)):
    if i==100:
        break
    actualValue = allChloroList1[i]
    predictedValue = regression1[0]*allTempList1[i]+regression1[1]
    residualValue = actualValue-predictedValue
    #STORE RESIDUAL VALUES IN ARRAY
    residualsArray1.append(residualValue)
                </code></pre>

<pre class="r"><code>#DETERMINE X AXIS FOR RESIDUAL DATA
timeseriesArray = []
timeseries1A = timeseries1.tolist()
for i in range(len(timeseries1A)):
    if i==100:
        break
    timeseriesArray.append(timeseries1A[i])

residuals1 = np.array(residualsArray1)
timeseries1 = np.array(timeseriesArray)
                </code></pre>

<pre class="r"><code>#PERFORM HARMONIC REGRESSION ON RESIDUALS
countCrosses = 0
firstvalue = residuals1[0]
mean = np.mean(residuals1)
for value in range(1, len(residuals1)):
    if firstvalue<mean and residuals1[value]>mean:
        countCrosses+=1
    firstvalue = residuals1[value]
print (countCrosses)
guess_freq = 2*np.pi*((countCrosses)/len(residuals1))
guess_amplitude = 3*np.std(residuals1)/(2**0.5)
guess_phase = 0
guess_offset = np.mean(residuals1)
p0=[guess_freq, guess_amplitude, guess_phase, guess_offset]
def my_sin(x, freq, amplitude, phase, offset):
    return np.sin(freq * x + phase) * amplitude + offset
fit = curve_fit(my_sin, timeseries1, residuals1, p0=p0)
#DATA_FIT IS SIN WAVE MODEL FOR HARMONIC REGRESSION ON RESIDUALS
data_fit = my_sin(timeseries1, *fit[0])

#start comment here
n=len(residuals1)
result = correlate(residuals1[-(n-100):], residuals1[n-100:], mode='full')

period = 0
firstvalue = result[0]
mean = np.mean(result)
firstplace = -1
for value in range(1, len(result)):
    if firstvalue<mean and result[value]>mean:
        if firstplace == -1:
            firstplace = value
        else:
            period = value - firstplace
            break
    firstvalue = result[value]

frequency = 1/(period)
guess_amplitude = 3*np.std(residuals1)/(2**0.5)
guess_phase = 0
guess_offset = np.mean(residuals1)
p0=[guess_amplitude, guess_phase, guess_offset]

def my_sin(x, amplitude, phase, offset):
    return np.sin(frequency * x + phase) * amplitude + offset

fit = curve_fit(my_sin, timeseries1, residuals1, p0=p0)
data_fit = my_sin(timeseries1, *fit[0])
                </code></pre>

<pre class="r"><code>#PLOT HARMONIC REGRESSION DATA ON TOP OF ACTUAL RESIDUAL VALUES
line_up, = plt.plot(data_fit, color = 'blue', label='Regression', linewidth='2')
line_down, = plt.plot(timeseries1, residuals1, label='Actual Residuals', linewidth='2')
plt.legend(handles=[line_up, line_down])
#plt.plot(result, color='red')
plt.xlabel("Time (Months, Only 100 For Sake of Visualization)")
plt.ylabel("Residual Values (mg/m^3)")
plt.suptitle("Residual Time Series of Scatter Plot")
plt.figure()
                </code></pre>


                <br><img src="img/linreg.png"/><br><br>
                <br><img src="img/regression.png"/><br><br>

                </section>

            <section id="appendix">
            <h2 class="sectionhead">Part 3: Data Analysis and Conclusions</h2>

             <p>After speaking with our mentors, we discovered that one possibility may be that each "side" of the triangle shape that the graph was forming was one period in each year; the correlation spiraled around over each year, and ultimately formed a triangle-shaped region. Furthermore, in regions where the chlorophyll levels were higher (usually near coasts), this correlation was more varied, which would explain the graph being cut off at 0; although the data would still theoretically form a triangle, it was cut off because chlorophyll cannot go below 0. From these results, we started performing nonparametric analyses, specifically harmonic regressions, on the data. Since temperature and chlorophyll values alter over the year in a sort of sinusoidal pattern, harmonic regressions would allow us to view my data linearly, and remove variability caused by variation over seasons. </p>

<pre class="r"><code>#DETERMINE ACCURACY OF MODEL
#CALCULATE PREDICTIONS THROUGH REGRESSIONS, AND PLOT AGAINST ACTUAL VALUES
allPoints1 = []
for value in allTempList1:
    point = my_sin(value, *fit[0])
    point += regression1[0]*value + regression1[1]
    allPoints1.append(point)

plt.scatter(allTempList1, allChloroList1)
plt.scatter(allTempList1, allPoints1, color='r')

plt.show()
                </code></pre>


        <p>These regressions allowed us to achieve our goal of predicting values of Chlorophyll based on temperature values: to compute each Chlorophyll, value, we used the formula "Chlorophyll Level = Linear Regression Value + Error Value from Harmonic Regression". We confirmed the accuracy of this prediction by omitting certain points of Chlorophyll from our time series, and comparing our calculated predictions with these actual values.</p>


            </section>


                </div><!-- Content -->

                <div id='sidebar' class="hidden-xs hidden-sm col-lg-2 col-md-2"><!-- Sidebar -->

                    <ul id='featured-nav' class="nav nav-list featured-nav nav-stacked">


                        <li>
                            <ul class="fa-style"><!-- Font Awesome -->
                               <!-- <li><a href="https://github.com/CommerceDataService/tutorial_ms_powerbi"><i class="fa fa-file-archive-o fa-lg"></i></a></li>
                                <li><a href="https://github.com/CommerceDataService/tutorial_ms_powerbi"><i class="fa fa-file-code-o fa-lg"></i></a></li>
                                <li><a href="https://github.com/CommerceDataService/tutorial_ms_powerbi"><i class="fa fa-github-square fa-lg"></i></a></li>-->
                            </ul>
                        </li>
                        <li>

                        </li>
                        <li>
                            <ul class="fa-style"><!-- Font Awesome -->
                                <!--<li><a href="https://github.com/CommerceDataService/tutorial_ms_powerbi"><i class="fa fa-file-archive-o fa-lg"></i></a></li>
                                <li><a href="https://github.com/CommerceDataService/tutorial_ms_powerbi"><i class="fa fa-file-code-o fa-lg"></i></a></li>
                                <li><a href="https://github.com/CommerceDataService/tutorial_ms_powerbi" title="Github Repo for MS Power BI Part 1"><i class="fa fa-github-square fa-lg"></i></a></li>-->
                            </ul>
                        </li>
                        <li>
                            <ul class="fa-style"><!-- Font Awesome -->
                                <li><a href="https://github.com/CommerceDataService/" title="Github Repo for Tutorial"><i class="fa fa-github-square fa-lg"></i></a></li>
                                <!--<li><a href=""><i class="fa fa-linkedin-square fa-lg"></i></a></li>
                                <li><a href=""><i class="fa fa-facebook-square fa-lg"></i></a></li>-->
                            </ul>
                        </li>
                        <li><a href="#intro">INTRODUCTION</a></li>
                        <li><a href="#data">DATA IS A PUBLIC GOOD</a></li>
                        <li><a href="#code">GETTING STARTED</a></li>
                        <li><a href="#visual">PART 1: PRELIMINARIES</a></li>
                        <li><a href="#conclusion">PART 2: UNDERSTANDING THE DATA</a></li>
                        <li><a href="#appendix">PART 3: DATA ANALYSIS AND CONCLUSION</a></li>

                    </ul>
                </div><!-- Sidebar -->
            </div><!-- Row -->
        </div><!-- Container-fluid -->
    </div>

</body>
</html>