change figure sizes

closed_cmr.qmd

@@ -47,7 +47,7 @@ plt.style.use('fivethirtyeight')
 plt.rcParams['axes.facecolor'] = 'white'
 plt.rcParams['figure.facecolor'] = 'white'
 # pal = sns.color_palette('Set2')
-# sns.set_palette(pal)ß
+# sns.set_palette(pal)
 
 # hyperparameters 
 SEED = 808
@@ -87,18 +87,18 @@ def get_history():
     col_labs = [f'y{t}' for t in range(1, 5)]
 
     # subset each dataset before augmenting 
-    is_red = sal_data.spp == 0
-    red = sal_data.loc[is_red, col_labs].to_numpy()
-    
     is_pyg = sal_data.spp == 1
+    is_red = sal_data.spp == 0
+
     pyg = sal_data.loc[is_pyg, col_labs].to_numpy()
+    red = sal_data.loc[is_red, col_labs].to_numpy()
 
     # augment each set separately since they differ in length
-    red_augmented = augment_history(red)
     pyg_augmented = augment_history(pyg)
+    red_augmented = augment_history(red)
 
     # recombine into one history 
-    history = np.concatenate((red_augmented, pyg_augmented))
+    history = np.concatenate((pyg_augmented, red_augmented))
 
     return history
 
@@ -121,7 +121,7 @@ In this example, I combine the two species into one `pm.Model` object, making us
 species_idx = np.repeat([0, 1], M)
 
 # coordinates identifying parameter each species  
-coords = {'species': ['red_cheeked', 'pygmy']}
+coords = {'species': ['pygmy', 'red_cheeked']}
 
 with pm.Model(coords=coords) as M0:
 
@@ -150,7 +150,7 @@ with M0:
 #| fig-cap: Traceplots for the salamander $M_0$ model. The red-cheeked salamander is in blue while the pygmy salamander is in red.
 #| label: fig-trace
 
-az.plot_trace(M0_idata);
+az.plot_trace(M0_idata, figsize=(10,4));
 ```
 
 For faster sampling, it's better to seperate the two species into two seperate models. On my machine, the individual species models finish sampling in 2-3 seconds, compared to 15-20 seconds for the two species model. That said, the two species model is somewhat more convenient.
@@ -174,7 +174,7 @@ Now we can look at the traceplots, this time excluding $\psi$ and looking at $N.
 ```{python}
 #| fig-cap: 'Traceplots for the salamander $M_0$ model, including $N$. Estimates from @hooten2019 are shown by the vertical lines'
 N_hooten = [229.6, 450.9]
-az.plot_trace(M0_idata, var_names=['p', 'N'], combined=True,
+az.plot_trace(M0_idata, var_names=['p', 'N'], combined=True, figsize=(10,4),
               lines=[("N", {}, [N_hooten])]);
 ```
 
@@ -192,7 +192,7 @@ N_samps = np.vstack(post.N.to_numpy())
 fig, ax = plt.subplots(1, 1, figsize=(4, 4))
 
 # add the scatter for each species
-labs = ['Red-backed','Pygmy']
+labs = ['Pygmy', 'Red-backed']
 ax.scatter(p_samps[:,0], N_samps[:,0], s=5, alpha=0.2, label=labs[0])
 ax.scatter(p_samps[:,1], N_samps[:,1], s=5, alpha=0.2, label=labs[1])
 

dist.qmd

@@ -22,6 +22,8 @@ The idea with distance sampling, also known as line-transect sampling, is that a
 Following @hooten2019, Chapter 24 and @royle2008, Chapter 7, I use the impala data from @burnham1980, who credits P. Hemingway with the dataset. In this dataset, 73 impalas were observed along a 60km transect. The distance values below are the perpendicular distances, in meters, from the transect. 
 
 ```{python}
+#| fig-cap: Histogram of the number of detected impalas at varying distances.
+#| label: fig-hist
 import pymc as pm
 import pytensor.tensor as pt
 import matplotlib.pyplot as plt
@@ -110,6 +112,7 @@ with distance:
 
 az.plot_trace(
     distance_idata, 
+    figsize=(10,4),
     var_names=['psi', 'sigma']
 );
 ```

docs/dist.html

@@ -316,7 +316,7 @@ <h1 class="title">Distance sampling</h1>
 <p>In this notebook, I explore how to fit distance sampling models for estimating the size of a closed population. Similar to the occupancy and closed capture-recapture notebooks, I use parameter-expanded data-augmentation (PX-DA) and the zero-inflated binomial model in this notebook.</p>
 <p>The idea with distance sampling, also known as line-transect sampling, is that a surveyer traverses a transect, typically in a boat or a plane. As they survey, they note when they detect an individual, or a group, from the species of interest, and further note the distance from the transect to the animal. Further, they note the angle to the animal(s), such that they can calculate the perpendicular distance from the animal to the transect. We assume that probability of detecting an animal <span class="math inline">\(p\)</span> decreases monotonically as the distance from the transect grows, e.g., <span class="math inline">\(p=\exp(-x^2/\sigma^2),\)</span> where <span class="math inline">\(x\)</span> is the distance and <span class="math inline">\(\sigma\)</span> is a scale parameter to be estimated. These simple assumptions permit the estimation of the population size <span class="math inline">\(N\)</span> as well as density <span class="math inline">\(D.\)</span></p>
 <p>Following <span class="citation" data-cites="hooten2019">Hooten and Hefley (<a href="#ref-hooten2019" role="doc-biblioref">2019</a>)</span>, Chapter 24 and <span class="citation" data-cites="royle2008">Royle and Dorazio (<a href="#ref-royle2008" role="doc-biblioref">2008</a>)</span>, Chapter 7, I use the impala data from <span class="citation" data-cites="burnham1980">Burnham, Anderson, and Laake (<a href="#ref-burnham1980" role="doc-biblioref">1980</a>)</span>, who credits P. Hemingway with the dataset. In this dataset, 73 impalas were observed along a 60km transect. The distance values below are the perpendicular distances, in meters, from the transect.</p>
-<div id="724f5c4d" class="cell" data-execution_count="1">
+<div id="cell-fig-hist" class="cell" data-execution_count="1">
 <div class="sourceCode cell-code" id="cb1"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> pymc <span class="im">as</span> pm</span>
 <span id="cb1-2"><a href="#cb1-2" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> pytensor.tensor <span class="im">as</span> pt</span>
 <span id="cb1-3"><a href="#cb1-3" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> matplotlib.pyplot <span class="im">as</span> plt</span>
@@ -361,15 +361,20 @@ <h1 class="title">Distance sampling</h1>
 <span id="cb1-42"><a href="#cb1-42" aria-hidden="true" tabindex="-1"></a></span>
 <span id="cb1-43"><a href="#cb1-43" aria-hidden="true" tabindex="-1"></a>plt.show()</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
 <div class="cell-output cell-output-display">
-<div>
-<figure class="figure">
-<p><img src="dist_files/figure-html/cell-2-output-1.png" width="429" height="408" class="figure-img"></p>
+<div id="fig-hist" class="quarto-figure quarto-figure-center quarto-float anchored">
+<figure class="quarto-float quarto-float-fig figure">
+<div aria-describedby="fig-hist-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
+<img src="dist_files/figure-html/fig-hist-output-1.png" width="429" height="408" class="figure-img">
+</div>
+<figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-hist-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
+Figure&nbsp;1: Histogram of the number of detected impalas at varying distances.
+</figcaption>
 </figure>
 </div>
 </div>
 </div>
 <p>Again, we treat this as a zero-inflated binomial model using PX-DA. The trick for doing so is to create a binary vector of length <span class="math inline">\(M\)</span>, <span class="math inline">\(y,\)</span> that represents whether the individual was detected during the study. Then, combine the indicator with the distance vector <span class="math inline">\(x\)</span> to create a the full dataset <span class="math inline">\((x,y).\)</span></p>
-<div id="659cd6be" class="cell" data-execution_count="2">
+<div id="14da023c" class="cell" data-execution_count="2">
 <div class="sourceCode cell-code" id="cb2"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a>n <span class="op">=</span> <span class="bu">len</span>(x_observed)</span>
 <span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a>unobserved_count <span class="op">=</span> M <span class="op">-</span> n</span>
 <span id="cb2-3"><a href="#cb2-3" aria-hidden="true" tabindex="-1"></a>zeros <span class="op">=</span> np.zeros(unobserved_count)</span>
@@ -399,21 +404,21 @@ <h1 class="title">Distance sampling</h1>
 <img src="dist_files/figure-html/fig-dist-output-1.svg" class="img-fluid figure-img">
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-dist-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-Figure&nbsp;1: Visual representation of the distance sampling model.
+Figure&nbsp;2: Visual representation of the distance sampling model.
 </figcaption>
 </figure>
 </div>
 </div>
 </div>
-<div id="f55db106" class="cell" data-execution_count="4">
+<div id="6d248158" class="cell" data-execution_count="4">
 <div class="sourceCode cell-code" id="cb4"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="cf">with</span> distance:</span>
 <span id="cb4-2"><a href="#cb4-2" aria-hidden="true" tabindex="-1"></a>    distance_idata <span class="op">=</span> pm.sample()</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
 <div class="cell-output cell-output-stderr">
 <pre><code>Auto-assigning NUTS sampler...
 Initializing NUTS using jitter+adapt_diag...
 Multiprocess sampling (4 chains in 4 jobs)
 NUTS: [psi, sigma, x_unobserved]
-Sampling 4 chains for 1_000 tune and 1_000 draw iterations (4_000 + 4_000 draws total) took 12 seconds.</code></pre>
+Sampling 4 chains for 1_000 tune and 1_000 draw iterations (4_000 + 4_000 draws total) took 11 seconds.</code></pre>
 </div>
 <div class="cell-output cell-output-display">
 
@@ -445,16 +450,17 @@ <h1 class="title">Distance sampling</h1>
 <div id="cell-fig-trace" class="cell" data-execution_count="5">
 <div class="sourceCode cell-code" id="cb6"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a>az.plot_trace(</span>
 <span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a>    distance_idata, </span>
-<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a>    var_names<span class="op">=</span>[<span class="st">'psi'</span>, <span class="st">'sigma'</span>]</span>
-<span id="cb6-4"><a href="#cb6-4" aria-hidden="true" tabindex="-1"></a>)<span class="op">;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
+<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a>    figsize<span class="op">=</span>(<span class="dv">10</span>,<span class="dv">4</span>),</span>
+<span id="cb6-4"><a href="#cb6-4" aria-hidden="true" tabindex="-1"></a>    var_names<span class="op">=</span>[<span class="st">'psi'</span>, <span class="st">'sigma'</span>]</span>
+<span id="cb6-5"><a href="#cb6-5" aria-hidden="true" tabindex="-1"></a>)<span class="op">;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
 <div class="cell-output cell-output-display">
 <div id="fig-trace" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-trace-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<img src="dist_files/figure-html/fig-trace-output-1.png" width="1021" height="367" class="figure-img">
+<img src="dist_files/figure-html/fig-trace-output-1.png" width="854" height="367" class="figure-img">
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-trace-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-Figure&nbsp;2: Traceplots for the distance sampling model.
+Figure&nbsp;3: Traceplots for the distance sampling model.
 </figcaption>
 </figure>
 </div>
@@ -509,7 +515,7 @@ <h1 class="title">Distance sampling</h1>
 <img src="dist_files/figure-html/fig-posteriors-output-1.png" width="766" height="382" class="figure-img">
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-posteriors-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-Figure&nbsp;3: Posterior distributions for <span class="math inline">\(N\)</span> and <span class="math inline">\(\sigma.\)</span> Estimates from <span class="citation" data-cites="royle2008">Royle and Dorazio (<a href="#ref-royle2008" role="doc-biblioref">2008</a>)</span> are shown with vertical lines.
+Figure&nbsp;4: Posterior distributions for <span class="math inline">\(N\)</span> and <span class="math inline">\(\sigma.\)</span> Estimates from <span class="citation" data-cites="royle2008">Royle and Dorazio (<a href="#ref-royle2008" role="doc-biblioref">2008</a>)</span> are shown with vertical lines.
 </figcaption>
 </figure>
 </div>
@@ -546,7 +552,7 @@ <h1 class="title">Distance sampling</h1>
 <img src="dist_files/figure-html/fig-func-output-1.png" width="498" height="408" class="figure-img">
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-func-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-Figure&nbsp;4: Posterior distribution for the detection function. The line represents the posterior mean while the shaded area is the 96% interval.
+Figure&nbsp;5: Posterior distribution for the detection function. The line represents the posterior mean while the shaded area is the 96% interval.
 </figcaption>
 </figure>
 </div>