| 1 |
{ |
| 2 |
"cells": [ |
| 3 |
{ |
| 4 |
"cell_type": "markdown", |
| 5 |
"metadata": {}, |
| 6 |
"source": [ |
| 7 |
"# Local Multi-Variable Moran Statistics" |
| 8 |
] |
| 9 |
}, |
| 10 |
{ |
| 11 |
"cell_type": "markdown", |
| 12 |
"metadata": {}, |
| 13 |
"source": [ |
| 14 |
"Local Moran statistics are very useful to assess spatial clusters (or outliers) in geographic data. Moran-style statistics are fundamentally based on the *covariance* of an outcome $y_i$ with other observations $y_j$, weighted according to some function that describes how near $i$ is to $j$. Classically, we describe that weight as $w_{ij}$, and collect that into a big matrix, $\\mathbf{W}$, that describes the relations between each site and every other site. \n", |
| 15 |
"\n", |
| 16 |
"The following discussion is a significantly condensed presentation of that found in [Wolf (2024)](https://doi.org/10.1080/24694452.2024.2326541) pages 1217-1225. \n", |
| 17 |
"\n", |
| 18 |
"The classical Moran statistic is stated as the relationship between $y$ and its surroundings. I'm assuming that $y$ is unit standardized, so that it has a mean of zero and a standard deviation of 1. Further, I'm assuming that our weights matrix is row-standardized with a diagonal of zero, meaning that $\\sum_j w_{ij} =1$ and $w_{ii}=0$. Further, this means that $\\sum_j w_{ij}y_j$ corresponds to the *spatially-weighted average* $y_j$ around $i$. With this understanding, the global $\\hat{I}$ estimator is often stated as: \n", |
| 19 |
"\n", |
| 20 |
"$$ \\hat{I} = \\frac{1}{n} \\sum_i y_i \\sum_j w_{ij}y_j $$\n", |
| 21 |
"\n", |
| 22 |
"You can understand this also as a kind of least squares estimator, arising from the following regression: \n", |
| 23 |
"\n", |
| 24 |
"$$ \\mathbf{W}y = Iy + \\epsilon \\ \\ \\ \\ \\ \\epsilon \\sim \\mathcal{N}(0,\\sigma^2) $$\n", |
| 25 |
"\n", |
| 26 |
"In this framing, we can think of the $I$ statistic in vector form as: \n", |
| 27 |
"\n", |
| 28 |
"$$ \\hat{I} = (y'y)^{-1}y'\\mathbf{W}y $$\n", |
| 29 |
"\n", |
| 30 |
"with the standardization we've used above. The *local* version of this statistic simply \"stops\" the outer summation over $i$, resulting in $n$ statistics overall:\n", |
| 31 |
"\n", |
| 32 |
"$$ \\hat{I}_i = \\frac{1}{n} y_i \\sum_j w_{ij} y_j $$\n", |
| 33 |
"\n", |
| 34 |
"This is equivalent to \"stopping\" the inner product between $y$ and $\\mathbf{W}y$ in the vector form, turning it into an element-wise product (spelled $\\circ$ in math below):\n", |
| 35 |
"\n", |
| 36 |
"$$ \\hat{I}_i = (y'y)^{-1}(y \\circ \\mathbf{W}y)$$\n", |
| 37 |
"\n", |
| 38 |
"This \"incomplete summation\" or \"inner-to-elementwise trick\" is how most of the local statistics are obtained from a global measure of covariance-based association. Any time you take an operation relating all pairs of observations, and sum that (or average that) over all the observations, you can create a \"local\" measure by just stopping that summation. \n", |
| 39 |
"\n", |
| 40 |
"## How do we introduce another variables? \n", |
| 41 |
"\n", |
| 42 |
"Often, it's useful to account for the spatial co-variation between two variables. Indeed, a common question is whether the *spatial pattern* of $y$ is similar to that of a second variable, $x$. Past attempts to link two variables like this include the Wartenberg statistic:\n", |
| 43 |
"\n", |
| 44 |
"$$ \\hat{I}_{xy,i} = x_i \\sum_j w_{ij}y_j $$\n", |
| 45 |
"$$ \\hat{\\mathbf{I}}_{xy} = x \\circ \\mathbf{W}y $$\n", |
| 46 |
"\n", |
| 47 |
"which relates $x_i$ to the local average of $y_j$ nearby. This is useful because it tells you whether a smoothed surface of $y$ looks like the surface of $x$. Perhaps less useful is the Lee (2001) innovation on the statistic, which seeks to compare the two smoothed patterns: \n", |
| 48 |
"\n", |
| 49 |
"$$ \\hat{L}_i = (\\sum_j w_{ij}x_j) * (\\sum_j w_{ij}y_j) $$\n", |
| 50 |
"$$ \\hat{\\mathbf{L}} = (\\mathbf{W}x) \\circ (\\mathbf{W}y) $$\n", |
| 51 |
"\n", |
| 52 |
"Both of these statistics are useful in their own ways, but generally are not able to separate out $x$'s influence on $y$ from $y$'s internal patterning over $\\mathbf{W}$. Instead, we're stuck making pairwise comparisons across $x$ and $y$ through $\\mathbf{W}$. \n", |
| 53 |
"\n", |
| 54 |
"## How can we introduce another *exogenous* variable? \n", |
| 55 |
"\n", |
| 56 |
"Indeed, it's often the case that $x$ represents some other factor we know is associated with $y$, but which cannot wholly explain $y$. When we just need a global statistic characterizing this relationship, the best solution is to use a spatial model, such as the spatial lag or spatial error models in `spreg`. Creating a local statistic from these models is difficult. Indeed, these models cannot be solved using ordinary least squares, so their non-linear matrix product cannot be \"stopped\" in the same way as above to create to an element-wise product. While LeSage (2009)'s tools can be used to examine the direct and indirect components of $\\beta$, which measure the effect that an intervention on $x$ at a given site at the site of the intervention on average (or, for indirect effects, *surrounding the intervention*), this is not itself a measure of the covariance between $x$ and $y$, nor can it measure the self-similarity of $y$ over $\\mathbf{W}$.\n", |
| 57 |
"\n", |
| 58 |
"Instead, what we often want is to \"remove\" or \"control\" for $x$. This \"adjusts\" $y$ for $x$, and then we can see the spatial pattern in this adjusted $y$. Think, for instance, of spatial clustering in house prices given information on house size. House size is decidedly spatially non-random and also impacts the sale price of a house. While we could develop a price-per-square-foot variable that reflects an adjustment to the house price based on the house size, this cannot scale to multiple factors. \n", |
| 59 |
"\n", |
| 60 |
"Another common approach is to to \"remove\" $x$ from $y$ using a regression. We regress $x$ on $y$ and analyze this model's residuals as if it they were our outcome of interest. That is, we do the first regression: \n", |
| 61 |
"\n", |
| 62 |
"$$ y = x\\beta + \\epsilon $$\n", |
| 63 |
"\n", |
| 64 |
"using a standard least squares estimator: \n", |
| 65 |
"\n", |
| 66 |
"$$ \\hat{\\beta} = (x'x)^{-1}x'y $$\n", |
| 67 |
"\n", |
| 68 |
"And use that to build a *residual* $y$, having removed its association with $x$: \n", |
| 69 |
"\n", |
| 70 |
"$$ e = y - x (x'x)^{-1} x' y $$\n", |
| 71 |
"\n", |
| 72 |
"and then use *this* in a Moran-form regression: \n", |
| 73 |
"\n", |
| 74 |
"$$ \\mathbf{W}e = I_{x\\rightarrow y}e + \\nu $$\n", |
| 75 |
"\n", |
| 76 |
"Here, the search for structure in $e$ has already assumed that variation in $y$ is fully \"caused by\" $x$ *first*. That is, the spatial structure of $y$ that matters is that which is fully independent of $x$'s pattern. This approach is similar to [Restricted Spatial Regression](https://www.tandfonline.com/doi/full/10.1080/01621459.2020.1788949), but proceeds very differently (and thus avoids its counter-intuitive effects). Note that $e = (I-P)y$ in the RSR framing. \n", |
| 77 |
"\n", |
| 78 |
"If we estimate this again using the same inner-to-elementwise trick, we can obtain: \n", |
| 79 |
"\n", |
| 80 |
"$$ I_{x\\rightarrow y} = e \\circ \\mathbf{W}e (e'e)^{-1}$$\n", |
| 81 |
"\n", |
| 82 |
"Assuming that there is only one variable $x$, this can be re-stated in terms of $y$ and $x$ with a little algebra: \n", |
| 83 |
"\n", |
| 84 |
"$$ I_{x\\rightarrow y} = \\left[ y\\circ \\mathbf{W}y - \\rho_{xy} x\\circ \\mathbf{W}y - \\rho_{xy} y\\circ\\mathbf{W}x + \\rho^2 x\\circ\\mathbf{W}x \\right] \\frac{n-1}{n} \\frac{1}{1-\\rho_{xy}^2}$$\n", |
| 85 |
"\n", |
| 86 |
"In this expression, you can see the Moran's $I$ all the way on the left of the bracketed expression, with two Wartenberg $I$s in the middle, and another Moran's $I$ for $x$ on the end. Each term is also scaled by $\\rho_{xy}$, depending on the directness of the relationship with $y$. You can think of this as $I$, rescaled by a measure of *indirect covariation* between $y$ and $x$ through $\\mathbf{W}$. Graphically, this is represented in the right facet of the following diagram:\n", |
| 87 |
"\n", |
| 88 |
"" |
| 89 |
] |
| 90 |
}, |
| 91 |
{ |
| 92 |
"cell_type": "markdown", |
| 93 |
"metadata": {}, |
| 94 |
"source": [ |
| 95 |
"Since we've assumed that all of the variation in $y$ \"comes from\" $x$ first, we must consider all of the paths through *any* $x$ into $y$. You can see that the correction terms that \"flow through\" $x$ must be corrected by $\\rho_{xy}$ each time. So, for the $I_{x \\rightarrow y}$ on the far right of the diagram, we have four possible paths between outcome $y_1$ and nearby avearge outcomes $\\mathbf{W}y_1$ at site $1$:\n", |
| 96 |
"\n", |
| 97 |
"1. The most direct path is from $y_1$ to $\\mathbf{W}y_1$, classic $I$\n", |
| 98 |
"2. Another indirect path goes from $y_1$ through $x_1$, then to $\\mathbf{W}y_1$, adjusting the Wartenberg $I$ by $\\rho$.\n", |
| 99 |
"3. A similar indirect path goes from $y_1$ directly to nearby $x_i$, and then to $\\mathbf{W}y_1$, resulting in the reversed Wartenberg estimator scaled by $\\rho$\n", |
| 100 |
"4. Another indirect path goes from $y_i$, through $x$, then $\\mathbf{W}x_1$, and then to $\\mathbf{W}y_1$, involving $\\rho$ twice and Moran's $I$ within $x$ alone.\n", |
| 101 |
"\n", |
| 102 |
"Hence, terms 2 and 3 are represented by the green path and the red+orange path, while term 4 corresponds to the red+maroon path. Terms 2 and 3 are Wartenberg-style bivariate Moran statistics corrected by $\\rho_{xy}$, while term 4 is a univariate Moran statistic for $x$, corrected by $\\rho_{xy}^2$.\n", |
| 103 |
"\n", |
| 104 |
"This may look like a bit of an **over-correction**, though! Typically, when we backtrack through a path and correct for the resulting term, we can \"delete\" those paths from the diagram. Once we backtrack through route 1,2, and 3, we have to *use the red $y_1 \\rightarrow x_1$ link twice* to walk backwards along route 4. This double-counts $\\rho_{xy}$, and thus over-corrects for the indirect association between $y$ and $x$. Indeed, this corresponds to assuming that *all of the variation in $y$ is assigned to $x$ *first*, and no variation in $y$ (nor $\\mathbf{W}y$} arises simultaneously with $x$. This is a very tight assumption indeed. For our housing example, this would *re-frame* the analysis. If we think of the house price as $y$, and the predicted house price given the house size $\\hat{y}$, then $e = y - \\hat{y}$ is the *premium* you'd pay (or discount you'd get) on a specific house of size $X=x$. When $y>\\hat{y}$, you're paying *more* for the house than typical for its size, and vice versa. Hence, this second regression *changes the analysis entirely* from house price to the *unique premium* paid for each property after modelling house price as a function of house size. Note, too, that this auxiliary regression strategy can use *any nonlinear model*, which makes it quite powerful indeed. \n", |
| 105 |
"\n", |
| 106 |
"But, what if we don't want to double-count the association between $y$ and $x$? What if we don't want to switch our analysis to talk about $e$? Can't we just *adjust* $y$ for its co-variance with $x$?\n", |
| 107 |
"\n", |
| 108 |
"Yes! We can use a Partial Moran statistic, which examines the \"direct\" path from $y$ to $\\mathbf{W}y$ and corrects it by the portion of variation due to $x$'s association with $y$, which drives a bit of the association between $y$ and $\\mathbf{W}y$:\n", |
| 109 |
"\n", |
| 110 |
"$$ \\mathbf{\\hat{I}}_{y|x} = \\left[y\\circ \\mathbf{W}y - \\rho_{xy} x\\circ\\mathbf{W}y \\right] \\frac{N-1}{N} \\frac{1}{1 - \\rho^2_{xy}}$$\n", |
| 111 |
"\n", |
| 112 |
"So, these both correspond to very basic \"corrections\" of the univariate $y$ using simple correlations, univariate Moran's $I$, or bivariate Moran's $I$. However, the method can be extended to more covariates, and its interpretation is easy to justify using some simple econometric theory (see the end of this notebook). In our housing example, this approach would be like *adjusting* the house price for its association with house size, but then continuing to analyse the adjusted house price, like a price index. The partial correction allows us to adjust for *a target house's characteristics* $x_i$ (not *all* nearby house's characteristics $\\mathbf{W}x_j$), in our measure of association.\n", |
| 113 |
"\n", |
| 114 |
"# Code time \n", |
| 115 |
"\n", |
| 116 |
"We'll start with a few core packages. First, these new statistics are implemented in the `moran_local_mv` module of the `esda` package. This stands for \"moran local multivariate\" statistic. In addition, we'll need to compare these stats to the standard Univariate Moran (`Moran_Local`) and bivariate/Wartenberg Moran (`Moran_Local_BV`) again from the `esda` package. We'll also want an easy way to calculate the correlation using `scipy.stats.pearsonr`, a way to read in our data (`geopandas`, `geodatasets`), a way to store the spatial relationships between observations, provided by the `libpysal.graph` module, and a visualization toolkit, `matplotlib.pyplot`. " |
| 117 |
] |
| 118 |
}, |
| 119 |
{ |
| 120 |
"cell_type": "code", |
| 121 |
"execution_count": 1, |
| 122 |
"metadata": {}, |
| 123 |
"outputs": [], |
| 124 |
"source": [ |
| 125 |
"from esda.moran_local_mv import (\n", |
| 126 |
" Partial_Moran_Local, Auxiliary_Moran_Local # new Moran\n", |
| 127 |
")\n", |
| 128 |
"from esda.moran import Moran_Local, Moran_Local_BV # standard Moran\n", |
| 129 |
"from libpysal import graph # construct spatial relations\n", |
| 130 |
"from scipy.stats import pearsonr # calculate correlation\n", |
| 131 |
"import geopandas, geodatasets # work with spatial data\n", |
| 132 |
"from matplotlib import pyplot as plt # visualize\n", |
| 133 |
"import seaborn # further visualization tools\n", |
| 134 |
"import numpy # math tools" |
| 135 |
] |
| 136 |
}, |
| 137 |
{ |
| 138 |
"cell_type": "markdown", |
| 139 |
"metadata": {}, |
| 140 |
"source": [ |
| 141 |
"We'll work with Airbnb data from Chicago, in order to understand how additional information $x$ informs our judgement about outcomes $y$. In this case, we'll consider the average price per head per night for airbnbs in neighborhoods of Chicago, and how it is informed by the overcrowding rate in those neighborhoods, which is a function itself of the population density and investment in housing stock in the neighborhood. Note, too, that price-per-head is *already correcting* the total booking price of an airbnb based on the number of people it can accommodate; we need a new technique to partition covariation correctly. " |
| 142 |
] |
| 143 |
}, |
| 144 |
{ |
| 145 |
"cell_type": "code", |
| 146 |
"execution_count": 14, |
| 147 |
"metadata": {}, |
| 148 |
"outputs": [], |
| 149 |
"source": [ |
| 150 |
"df = geopandas.read_file(\n", |
| 151 |
" geodatasets.get_path(\"geoda.airbnb\")\n", |
| 152 |
").dropna()" |
| 153 |
] |
| 154 |
}, |
| 155 |
{ |
| 156 |
"cell_type": "code", |
| 157 |
"execution_count": 15, |
| 158 |
"metadata": { |
| 159 |
"scrolled": true |
| 160 |
}, |
| 161 |
"outputs": [ |
| 162 |
{ |
| 163 |
"data": { |
| 164 |
"text/plain": [ |
| 165 |
"<Axes: >" |
| 166 |
] |
| 167 |
}, |
| 168 |
"execution_count": 15, |
| 169 |
"metadata": {}, |
| 170 |
"output_type": "execute_result" |
| 171 |
}, |
| 172 |
{ |
| 173 |
"data": { |
| 174 |
"image/png": "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", |
| 175 |
"text/plain": [ |
| 176 |
"<Figure size 640x480 with 1 Axes>" |
| 177 |
] |
| 178 |
}, |
| 179 |
"metadata": {}, |
| 180 |
"output_type": "display_data" |
| 181 |
} |
| 182 |
], |
| 183 |
"source": [ |
| 184 |
"df.plot(\"price_pp\")" |
| 185 |
] |
| 186 |
}, |
| 187 |
{ |
| 188 |
"cell_type": "markdown", |
| 189 |
"metadata": {}, |
| 190 |
"source": [ |
| 191 |
"In order to estimate the statistics, we need to get our data. We'll use the `price_pp` to get the price per person per night for Airbnbs in each neighborhood. And, we'll use the `crowded` column for the percentages of properties that are over-crowded. We'll build a simple contiguity weights matrix to explain the spatial adjacencies between neighborhoods. " |
| 192 |
] |
| 193 |
}, |
| 194 |
{ |
| 195 |
"cell_type": "code", |
| 196 |
"execution_count": 16, |
| 197 |
"metadata": {}, |
| 198 |
"outputs": [], |
| 199 |
"source": [ |
| 200 |
"y = df.price_pp.values.reshape(-1,1)\n", |
| 201 |
"x = df.crowded.values.reshape(-1,1)\n", |
| 202 |
"w = graph.Graph.build_contiguity(df).transform(\"R\")" |
| 203 |
] |
| 204 |
}, |
| 205 |
{ |
| 206 |
"cell_type": "markdown", |
| 207 |
"metadata": {}, |
| 208 |
"source": [ |
| 209 |
"Now, with these defined, we can calculate all of our statistics. First, we can note that each statistic corresponds to a different question of relation, like we discussed above: " |
| 210 |
] |
| 211 |
}, |
| 212 |
{ |
| 213 |
"cell_type": "code", |
| 214 |
"execution_count": 17, |
| 215 |
"metadata": {}, |
| 216 |
"outputs": [ |
| 217 |
{ |
| 218 |
"name": "stderr", |
| 219 |
"output_type": "stream", |
| 220 |
"text": [ |
| 221 |
"Simulating by site: 100%|██████████████████████████████████████████████████████████████████| 66/66 [00:00<00:00, 12136.62it/s]\n", |
| 222 |
"Simulating by site: 100%|██████████████████████████████████████████████████████████████████| 66/66 [00:00<00:00, 21919.71it/s]\n" |
| 223 |
] |
| 224 |
} |
| 225 |
], |
| 226 |
"source": [ |
| 227 |
"# are expensive places near expensive places?\n", |
| 228 |
"Iy = Moran_Local(y, w)\n", |
| 229 |
"# are crowded plcaes near crowded places?\n", |
| 230 |
"Ix = Moran_Local(x, w) \n", |
| 231 |
"# are expensive places also crowded places? \n", |
| 232 |
"rho = pearsonr(y.squeeze(),x.squeeze())[0]\n", |
| 233 |
"# are expensive places near crowded places?\n", |
| 234 |
"Ixy = Moran_Local_BV(x, y, w) \n", |
| 235 |
"# are crowded places near expensive places?\n", |
| 236 |
"Iyx = Moran_Local_BV(y, x, w)\n", |
| 237 |
"# are expensive places near expensive places, adjusted for crowding?\n", |
| 238 |
"p_lmo = Partial_Moran_Local().fit(x,y,w)\n", |
| 239 |
"# assuming crowding explains price, \n", |
| 240 |
"# are places with high post-crowding premiums near one another?\n", |
| 241 |
"a_lmo = Auxiliary_Moran_Local().fit(x,y,w)\n", |
| 242 |
"\n", |
| 243 |
"# this is a scaling factor for the multivariable statistics\n", |
| 244 |
"scaling = (w.n - 1)/(w.n) * (1/(1-rho**2))" |
| 245 |
] |
| 246 |
}, |
| 247 |
{ |
| 248 |
"cell_type": "markdown", |
| 249 |
"metadata": {}, |
| 250 |
"source": [ |
| 251 |
"You can then see that the multivariate estimators for the partial and auxiliary regressive versions of the Local Moran statistic are just combinations of these simple pairwise statistics when there is only one $x$ variable: " |
| 252 |
] |
| 253 |
}, |
| 254 |
{ |
| 255 |
"cell_type": "code", |
| 256 |
"execution_count": 18, |
| 257 |
"metadata": {}, |
| 258 |
"outputs": [ |
| 259 |
{ |
| 260 |
"data": { |
| 261 |
"image/png": "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", |
| 262 |
"text/plain": [ |
| 263 |
"<Figure size 640x480 with 2 Axes>" |
| 264 |
] |
| 265 |
}, |
| 266 |
"metadata": {}, |
| 267 |
"output_type": "display_data" |
| 268 |
} |
| 269 |
], |
| 270 |
"source": [ |
| 271 |
"f,ax = plt.subplots(1,2, subplot_kw=dict(aspect='equal'))\n", |
| 272 |
"\n", |
| 273 |
"ax[0].scatter(\n", |
| 274 |
" (Iy.Is - rho * Ixy.Is)*scaling, \n", |
| 275 |
" p_lmo.associations_\n", |
| 276 |
")\n", |
| 277 |
"\n", |
| 278 |
"ax[1].scatter(\n", |
| 279 |
" (Iy.Is - rho*Ixy.Is - rho*Iyx.Is + rho**2*Ix.Is)*scaling,\n", |
| 280 |
" a_lmo.associations_)\n", |
| 281 |
"\n", |
| 282 |
"for i in range(2):\n", |
| 283 |
" xmin, xmax = ax[i].get_xlim()\n", |
| 284 |
" ymin, ymax = ax[i].get_ylim()\n", |
| 285 |
" left = min([xmin, ymin])\n", |
| 286 |
" right = max([xmax, ymax])\n", |
| 287 |
" ax[i].plot([left, right], [left, right], color='k', linestyle=\":\")\n", |
| 288 |
" ax[i].set_xlim(left, right)\n", |
| 289 |
" ax[i].set_ylim(left, right)\n", |
| 290 |
" if i==0:\n", |
| 291 |
" ax[i].set_ylabel(\"Splitting an OLS\")\n", |
| 292 |
" ax[i].set_xlabel(\"Combining basic stats\")\n", |
| 293 |
" ax[i].set_title(['Partial Local Moran', 'Axiliary Local Moran'][i])\n", |
| 294 |
" seaborn.despine(ax=ax[i])" |
| 295 |
] |
| 296 |
}, |
| 297 |
{ |
| 298 |
"cell_type": "markdown", |
| 299 |
"metadata": {}, |
| 300 |
"source": [ |
| 301 |
"You can see the direction of changes shown below. The changes in the partial statistic can only move from left to right, since the correction is only applied to $y$, not $\\mathbf{W}y$. The component that \"corrects\" the $\\mathbf{W}y$ for $x$ comes from that due to $\\mathbf{W}x$, which arises in the auxiliary regression version, shown on the right. \n" |
| 302 |
] |
| 303 |
}, |
| 304 |
{ |
| 305 |
"cell_type": "code", |
| 306 |
"execution_count": 19, |
| 307 |
"metadata": { |
| 308 |
"jupyter": { |
| 309 |
"source_hidden": false |
| 310 |
} |
| 311 |
}, |
| 312 |
"outputs": [], |
| 313 |
"source": [ |
| 314 |
"#| echo: false\n", |
| 315 |
"def quadcolor(x,y, colors=None):\n", |
| 316 |
" off_sign = numpy.sign(y) != numpy.sign(x)\n", |
| 317 |
" neg_y = y < 0\n", |
| 318 |
" quad = off_sign + neg_y*2\n", |
| 319 |
" if colors is None:\n", |
| 320 |
" colors= numpy.array(['red','powderblue','steelblue','lightsalmon'])\n", |
| 321 |
" return colors[quad]" |
| 322 |
] |
| 323 |
}, |
| 324 |
{ |
| 325 |
"cell_type": "code", |
| 326 |
"execution_count": 20, |
| 327 |
"metadata": { |
| 328 |
"jupyter": { |
| 329 |
"source_hidden": false |
| 330 |
} |
| 331 |
}, |
| 332 |
"outputs": [ |
| 333 |
{ |
| 334 |
"data": { |
| 335 |
"image/png": "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", |
| 336 |
"text/plain": [ |
| 337 |
"<Figure size 800x400 with 2 Axes>" |
| 338 |
] |
| 339 |
}, |
| 340 |
"metadata": {}, |
| 341 |
"output_type": "display_data" |
| 342 |
} |
| 343 |
], |
| 344 |
"source": [ |
| 345 |
"#| echo: false\n", |
| 346 |
"f,ax = plt.subplots(1,2,figsize=(8,4))\n", |
| 347 |
"for i,ax_ in enumerate(ax):\n", |
| 348 |
" start_x, start_y = Iy.z, Iy.w.sparse @ Iy.z\n", |
| 349 |
" new_x, new_y = (p_lmo.partials_.T, a_lmo.partials_.T)[i]\n", |
| 350 |
" dx, dy = new_x - start_x, new_y - start_y\n", |
| 351 |
" colors = quadcolor(new_x, new_y)\n", |
| 352 |
" for j in range(Iy.w.n):\n", |
| 353 |
" ax_.arrow(\n", |
| 354 |
" x=start_x[j], y=start_y[j],\n", |
| 355 |
" dx = dx[j], dy = dy[j],\n", |
| 356 |
" width=.02, shape='full', \n", |
| 357 |
" length_includes_head=True, linewidth=0,\n", |
| 358 |
" head_width=.1, color=colors[j]\n", |
| 359 |
" )\n", |
| 360 |
" ax_.axvline(0, color='k', linestyle=':')\n", |
| 361 |
" ax_.axhline(0, color='k', linestyle=':')\n", |
| 362 |
" for j in range(2):\n", |
| 363 |
" for k in range(2):\n", |
| 364 |
" ax_.annotate(\n", |
| 365 |
" text=['Hotspot', 'Pit', 'Peak', 'Coldspot'][j+2*k],\n", |
| 366 |
" xy=(.1 if j else .8,\n", |
| 367 |
" .025 if k else .6\n", |
| 368 |
" ),\n", |
| 369 |
" xycoords=\"axes fraction\"\n", |
| 370 |
" )\n", |
| 371 |
" ax_.set_xlabel(r\"$y-\\rho x$\")\n", |
| 372 |
" ax_.set_ylabel((r\"$\\mathbf{W}y$\", r\"$\\mathbf{W}(y-\\rho x)$\")[i])\n", |
| 373 |
"f.tight_layout()\n", |
| 374 |
"plt.show()" |
| 375 |
] |
| 376 |
}, |
| 377 |
{ |
| 378 |
"cell_type": "markdown", |
| 379 |
"metadata": {}, |
| 380 |
"source": [ |
| 381 |
"Here, we've colored the statistics according to their *new* quadrant. In the same fashion as a regular Local Moran statistic, we get quadrants that describe whether $y-\\rho x$ is above or below the mean, (which, in both cases, is zero), and whether $\\mathbf{W}y$ (or $\\mathbf{W}(y - \\rho x)$) is above or below its mean (which, again, is zero by construction). Further, you can note the four-part quadrant breakdown is explained by the annotations:\n", |
| 382 |
"\n", |
| 383 |
"- A *hotspot* is an area where the observation under study is above its mean, and also above the surrounding mean.\n", |
| 384 |
"- A *pit* is an area where the observation under study is below its mean, but is surrounded by observations larger than the mean. \n", |
| 385 |
"- A *coldspot* is a hotspot in reverse: observations are smaller than the mean, and surrounded by values smaller than the mean. \n", |
| 386 |
"- A *peak* is a pit in reverse: the observation under study is above the mean, but is surrounded by values below the mean. \n", |
| 387 |
"\n", |
| 388 |
"In the case of the multivariable statistics, the interpretation remains the same; we just change what values we're comparing at the site and surrounding. For the partial Moran, we compare the adjusted $y-\\rho x$ to $\\mathbf{W}y$, while the full Auxiliary Regression Moran compares $y - \\rho x$ to $\\mathbf{W}y - \\rho \\mathbf{W}x$. Further, it might help to recognize that arrows in the left plot represent *only* the lateral (left-right, horizontal) component of arrows in the right plot. The left plot shows how information conditions our *site-specific* judgement, while the right plot shows how information conditions our judgement about the entire relationship. The vertical component of those vectors corresponds to the contribution of the $\\rho\\mathbf{W}x$ component of the y-axis term. \n", |
| 389 |
"\n", |
| 390 |
"We can also center the arrows on a common axis. This will let us see at a higher level whether information about crowding changes our judgement about the pattern of prices in Chicago. " |
| 391 |
] |
| 392 |
}, |
| 393 |
{ |
| 394 |
"cell_type": "code", |
| 395 |
"execution_count": 21, |
| 396 |
"metadata": { |
| 397 |
"jupyter": { |
| 398 |
"source_hidden": false |
| 399 |
} |
| 400 |
}, |
| 401 |
"outputs": [], |
| 402 |
"source": [ |
| 403 |
"#| echo: false\n", |
| 404 |
"def circular_hist(ax, x, bins=16, density=True, offset=0, gaps=True):\n", |
| 405 |
" \"\"\"\n", |
| 406 |
" Produce a circular histogram of angles on ax.\n", |
| 407 |
"\n", |
| 408 |
" Parameters\n", |
| 409 |
" ----------\n", |
| 410 |
" ax : matplotlib.axes._subplots.PolarAxesSubplot\n", |
| 411 |
" axis instance created with subplot_kw=dict(projection='polar').\n", |
| 412 |
"\n", |
| 413 |
" x : array\n", |
| 414 |
" Angles to plot, expected in units of radians.\n", |
| 415 |
"\n", |
| 416 |
" bins : int, optional\n", |
| 417 |
" Defines the number of equal-width bins in the range. The default is 16.\n", |
| 418 |
"\n", |
| 419 |
" density : bool, optional\n", |
| 420 |
" If True plot frequency proportional to area. If False plot frequency\n", |
| 421 |
" proportional to radius. The default is True.\n", |
| 422 |
"\n", |
| 423 |
" offset : float, optional\n", |
| 424 |
" Sets the offset for the location of the 0 direction in units of\n", |
| 425 |
" radians. The default is 0.\n", |
| 426 |
"\n", |
| 427 |
" gaps : bool, optional\n", |
| 428 |
" Whether to allow gaps between bins. When gaps = False the bins are\n", |
| 429 |
" forced to partition the entire [-pi, pi] range. The default is True.\n", |
| 430 |
"\n", |
| 431 |
" Returns\n", |
| 432 |
" -------\n", |
| 433 |
" n : array or list of arrays\n", |
| 434 |
" The number of values in each bin.\n", |
| 435 |
"\n", |
| 436 |
" bins : array\n", |
| 437 |
" The edges of the bins.\n", |
| 438 |
"\n", |
| 439 |
" patches : `.BarContainer` or list of a single `.Polygon`\n", |
| 440 |
" Container of individual artists used to create the histogram\n", |
| 441 |
" or list of such containers if there are multiple input datasets.\n", |
| 442 |
" \"\"\"\n", |
| 443 |
" np = numpy\n", |
| 444 |
" # Wrap angles to [-pi, pi)\n", |
| 445 |
" x = (x+np.pi) % (2*np.pi) - np.pi\n", |
| 446 |
"\n", |
| 447 |
" # Force bins to partition entire circle\n", |
| 448 |
" if not gaps:\n", |
| 449 |
" bins = np.linspace(-np.pi, np.pi, num=bins+1)\n", |
| 450 |
"\n", |
| 451 |
" # Bin data and record counts\n", |
| 452 |
" n, bins = np.histogram(x, bins=bins)\n", |
| 453 |
"\n", |
| 454 |
" # Compute width of each bin\n", |
| 455 |
" widths = np.diff(bins)\n", |
| 456 |
"\n", |
| 457 |
" # By default plot frequency proportional to area\n", |
| 458 |
" if density:\n", |
| 459 |
" # Area to assign each bin\n", |
| 460 |
" area = n / x.size\n", |
| 461 |
" # Calculate corresponding bin radius\n", |
| 462 |
" radius = (area/np.pi) ** .5\n", |
| 463 |
" # Otherwise plot frequency proportional to radius\n", |
| 464 |
" else:\n", |
| 465 |
" radius = n\n", |
| 466 |
"\n", |
| 467 |
" # Plot data on ax\n", |
| 468 |
" patches = ax.bar(bins[:-1], radius, zorder=1, \n", |
| 469 |
" align='edge', width=widths, linewidth=1\n", |
| 470 |
" )\n", |
| 471 |
"\n", |
| 472 |
" # Set the direction of the zero angle\n", |
| 473 |
" ax.set_theta_offset(offset)\n", |
| 474 |
"\n", |
| 475 |
" # Remove ylabels for area plots (they are mostly obstructive)\n", |
| 476 |
" if density:\n", |
| 477 |
" ax.set_yticks([])\n", |
| 478 |
"\n", |
| 479 |
" return n, bins, patches" |
| 480 |
] |
| 481 |
}, |
| 482 |
{ |
| 483 |
"cell_type": "markdown", |
| 484 |
"metadata": {}, |
| 485 |
"source": [ |
| 486 |
"For the partial Moran, we can visualize the changes along the $y$ axis as different \"arrows\" pushing either left or right. This is fine because the correction is only be applied on the $x$ component. For the auxiliary Moran, we need to build a Rose diagram that maps the typical move in each direction. This puts the original $I_i$ value at the center of the diagram, and shows the number of moves that go in each of the directions. Longer bars/pie slices indicate more moves in that direction. " |
| 487 |
] |
| 488 |
}, |
| 489 |
{ |
| 490 |
"cell_type": "code", |
| 491 |
"execution_count": 22, |
| 492 |
"metadata": { |
| 493 |
"jupyter": { |
| 494 |
"source_hidden": false |
| 495 |
} |
| 496 |
}, |
| 497 |
"outputs": [ |
| 498 |
{ |
| 499 |
"data": { |
| 500 |
"image/png": "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", |
| 501 |
"text/plain": [ |
| 502 |
"<Figure size 800x400 with 2 Axes>" |
| 503 |
] |
| 504 |
}, |
| 505 |
"metadata": {}, |
| 506 |
"output_type": "display_data" |
| 507 |
} |
| 508 |
], |
| 509 |
"source": [ |
| 510 |
"#| echo: false\n", |
| 511 |
"f = plt.figure(figsize=(8,4))\n", |
| 512 |
"ax1 = plt.subplot(1,2,1)\n", |
| 513 |
"ax2 = plt.subplot(1,2,2, projection='polar')\n", |
| 514 |
"for i,ax_ in enumerate((ax1, ax2)):\n", |
| 515 |
" start_x, start_y = Iy.z, Iy.w.sparse @ Iy.z\n", |
| 516 |
" new_x, new_y = (p_lmo.partials_.T, a_lmo.partials_.T)[i]\n", |
| 517 |
" dx, dy = new_x - start_x, new_y - start_y\n", |
| 518 |
" colors = quadcolor(dx, dy if i > 0 else new_y,\n", |
| 519 |
" colors=numpy.array(\n", |
| 520 |
" [\"forestgreen\", \n", |
| 521 |
" \"violet\", \n", |
| 522 |
" \"purple\", \n", |
| 523 |
" \"lightgreen\"\n", |
| 524 |
" ]\n", |
| 525 |
" )\n", |
| 526 |
" )\n", |
| 527 |
" if i == 0:\n", |
| 528 |
" for j in range(Iy.w.n):\n", |
| 529 |
" ax_.arrow(\n", |
| 530 |
" x=0, y=0 if i > 0 else start_y[j],\n", |
| 531 |
" dx = dx[j], dy = dy[j],\n", |
| 532 |
" width=.02, shape='full', \n", |
| 533 |
" length_includes_head=True, linewidth=0,\n", |
| 534 |
" head_width=.05, color=colors[j]\n", |
| 535 |
" )\n", |
| 536 |
" ax_.axvline(0, color='k', linestyle=':')\n", |
| 537 |
" ax_.axhline(0, color='k', linestyle=':')\n", |
| 538 |
" ax1.set_title(\"Change, Univariate to Part-Conditional\")\n", |
| 539 |
" seaborn.despine(ax=ax1)\n", |
| 540 |
" ax1.legend(\n", |
| 541 |
" handles = \n", |
| 542 |
" [\n", |
| 543 |
" plt.arrow(0,0,1,0,\n", |
| 544 |
" color=[\"forestgreen\", \"violet\", \"purple\", \"lightgreen\"][i]) for i in range(4)\n", |
| 545 |
" ],\n", |
| 546 |
" labels = [\n", |
| 547 |
" \"X Heats y\", \n", |
| 548 |
" \"X Pits y\", \n", |
| 549 |
" \"X Cools y\", \n", |
| 550 |
" \"X Peaks y\"\n", |
| 551 |
" ]\n", |
| 552 |
" )\n", |
| 553 |
" else:\n", |
| 554 |
" theta = numpy.arctan2(dy, dx)\n", |
| 555 |
" n, bins, patches = circular_hist(ax=ax2, x=theta, bins=16, gaps=False)\n", |
| 556 |
" for bin_, patch_ in zip(bins,patches.patches):\n", |
| 557 |
" if -numpy.pi/2 <= bin_ < 0:\n", |
| 558 |
" patch_.set_facecolor(\"lightgreen\")\n", |
| 559 |
" elif 0 <= bin_ < numpy.pi/2:\n", |
| 560 |
" patch_.set_facecolor(\"forestgreen\")\n", |
| 561 |
" elif numpy.pi/2 <= bin_ < numpy.pi:\n", |
| 562 |
" patch_.set_facecolor(\"violet\")\n", |
| 563 |
" elif -numpy.pi <= bin_ < -numpy.pi/2:\n", |
| 564 |
" patch_.set_facecolor(\"purple\")\n", |
| 565 |
" ax2.set_title(\"Change, Univariate to Full Conditional\")\n", |
| 566 |
"f.tight_layout()\n", |
| 567 |
"plt.show()" |
| 568 |
] |
| 569 |
}, |
| 570 |
{ |
| 571 |
"cell_type": "markdown", |
| 572 |
"metadata": {}, |
| 573 |
"source": [ |
| 574 |
"From the partial statistic, you can see that most of the moves occur in the hotspot and coldspot quadrants, but quite a few very large moves push things towards \"peaked\"ness. These are many of the moves on the bottom right quadrant, where even a few observations shift from being a coldspot to being a peak. This means that they're in relatively low-priced areas, given their crowding levels. I refer to this as \"x peaks y\", since it indicates that $x$ pushes the local statistic towards the \"peak\" quadrant. Note that this does *not* indicate a change in classification, which we can see in a second. \n", |
| 575 |
"\n", |
| 576 |
"For the full conditional case, you can see from these diagrams that the inclusion of the crowding information pushes most strongly in the cluster/outlier direction. This means that the additional information about crowding increases our perception of how the price-per-person-per-night clusters spatially. \n", |
| 577 |
"\n", |
| 578 |
"To see the actual table of class changes (ignoring statistical significance for now), we can use `pandas` crosstabs to build the following nice tables: " |
| 579 |
] |
| 580 |
}, |
| 581 |
{ |
| 582 |
"cell_type": "code", |
| 583 |
"execution_count": 23, |
| 584 |
"metadata": {}, |
| 585 |
"outputs": [], |
| 586 |
"source": [ |
| 587 |
"#| echo: false\n", |
| 588 |
"#| #| jupyter: {source_hidden: false}\n", |
| 589 |
"import pandas\n", |
| 590 |
"\n", |
| 591 |
"partial_moves = pandas.crosstab(\n", |
| 592 |
" Iy.q,\n", |
| 593 |
" p_lmo.labels_,\n", |
| 594 |
" )\n", |
| 595 |
"partial_moves.columns = partial_moves.index = ['hotspot', 'pit', 'coldspot', 'peak']\n", |
| 596 |
"partial_moves.columns.name = \"Partial\"\n", |
| 597 |
"partial_moves.index.name = \"Univariate\"\n", |
| 598 |
"\n", |
| 599 |
"aux_moves = pandas.crosstab(\n", |
| 600 |
" Iy.q,\n", |
| 601 |
" a_lmo.labels_,\n", |
| 602 |
" )\n", |
| 603 |
"aux_moves.columns = aux_moves.index = ['hotspot', 'pit', 'coldspot', 'peak']\n", |
| 604 |
"aux_moves.columns.name = \"Auxiliary\"\n", |
| 605 |
"aux_moves.index.name = \"Univariate\"" |
| 606 |
] |
| 607 |
}, |
| 608 |
{ |
| 609 |
"cell_type": "code", |
| 610 |
"execution_count": 24, |
| 611 |
"metadata": { |
| 612 |
"jupyter": { |
| 613 |
"source_hidden": false |
| 614 |
} |
| 615 |
}, |
| 616 |
"outputs": [ |
| 617 |
{ |
| 618 |
"data": { |
| 619 |
"text/html": [ |
| 620 |
"<div>\n", |
| 621 |
"<style scoped>\n", |
| 622 |
" .dataframe tbody tr th:only-of-type {\n", |
| 623 |
" vertical-align: middle;\n", |
| 624 |
" }\n", |
| 625 |
"\n", |
| 626 |
" .dataframe tbody tr th {\n", |
| 627 |
" vertical-align: top;\n", |
| 628 |
" }\n", |
| 629 |
"\n", |
| 630 |
" .dataframe thead th {\n", |
| 631 |
" text-align: right;\n", |
| 632 |
" }\n", |
| 633 |
"</style>\n", |
| 634 |
"<table border=\"1\" class=\"dataframe\">\n", |
| 635 |
" <thead>\n", |
| 636 |
" <tr style=\"text-align: right;\">\n", |
| 637 |
" <th>Partial</th>\n", |
| 638 |
" <th>hotspot</th>\n", |
| 639 |
" <th>pit</th>\n", |
| 640 |
" <th>coldspot</th>\n", |
| 641 |
" <th>peak</th>\n", |
| 642 |
" </tr>\n", |
| 643 |
" <tr>\n", |
| 644 |
" <th>Univariate</th>\n", |
| 645 |
" <th></th>\n", |
| 646 |
" <th></th>\n", |
| 647 |
" <th></th>\n", |
| 648 |
" <th></th>\n", |
| 649 |
" </tr>\n", |
| 650 |
" </thead>\n", |
| 651 |
" <tbody>\n", |
| 652 |
" <tr>\n", |
| 653 |
" <th>hotspot</th>\n", |
| 654 |
" <td>14</td>\n", |
| 655 |
" <td>3</td>\n", |
| 656 |
" <td>0</td>\n", |
| 657 |
" <td>0</td>\n", |
| 658 |
" </tr>\n", |
| 659 |
" <tr>\n", |
| 660 |
" <th>pit</th>\n", |
| 661 |
" <td>4</td>\n", |
| 662 |
" <td>9</td>\n", |
| 663 |
" <td>0</td>\n", |
| 664 |
" <td>0</td>\n", |
| 665 |
" </tr>\n", |
| 666 |
" <tr>\n", |
| 667 |
" <th>coldspot</th>\n", |
| 668 |
" <td>0</td>\n", |
| 669 |
" <td>0</td>\n", |
| 670 |
" <td>24</td>\n", |
| 671 |
" <td>3</td>\n", |
| 672 |
" </tr>\n", |
| 673 |
" <tr>\n", |
| 674 |
" <th>peak</th>\n", |
| 675 |
" <td>0</td>\n", |
| 676 |
" <td>0</td>\n", |
| 677 |
" <td>1</td>\n", |
| 678 |
" <td>8</td>\n", |
| 679 |
" </tr>\n", |
| 680 |
" </tbody>\n", |
| 681 |
"</table>\n", |
| 682 |
"</div>" |
| 683 |
], |
| 684 |
"text/plain": [ |
| 685 |
"Partial hotspot pit coldspot peak\n", |
| 686 |
"Univariate \n", |
| 687 |
"hotspot 14 3 0 0\n", |
| 688 |
"pit 4 9 0 0\n", |
| 689 |
"coldspot 0 0 24 3\n", |
| 690 |
"peak 0 0 1 8" |
| 691 |
] |
| 692 |
}, |
| 693 |
"execution_count": 24, |
| 694 |
"metadata": {}, |
| 695 |
"output_type": "execute_result" |
| 696 |
} |
| 697 |
], |
| 698 |
"source": [ |
| 699 |
"#| echo: false\n", |
| 700 |
"partial_moves" |
| 701 |
] |
| 702 |
}, |
| 703 |
{ |
| 704 |
"cell_type": "code", |
| 705 |
"execution_count": 25, |
| 706 |
"metadata": { |
| 707 |
"jupyter": { |
| 708 |
"source_hidden": false |
| 709 |
} |
| 710 |
}, |
| 711 |
"outputs": [ |
| 712 |
{ |
| 713 |
"data": { |
| 714 |
"text/html": [ |
| 715 |
"<div>\n", |
| 716 |
"<style scoped>\n", |
| 717 |
" .dataframe tbody tr th:only-of-type {\n", |
| 718 |
" vertical-align: middle;\n", |
| 719 |
" }\n", |
| 720 |
"\n", |
| 721 |
" .dataframe tbody tr th {\n", |
| 722 |
" vertical-align: top;\n", |
| 723 |
" }\n", |
| 724 |
"\n", |
| 725 |
" .dataframe thead th {\n", |
| 726 |
" text-align: right;\n", |
| 727 |
" }\n", |
| 728 |
"</style>\n", |
| 729 |
"<table border=\"1\" class=\"dataframe\">\n", |
| 730 |
" <thead>\n", |
| 731 |
" <tr style=\"text-align: right;\">\n", |
| 732 |
" <th>Auxiliary</th>\n", |
| 733 |
" <th>hotspot</th>\n", |
| 734 |
" <th>pit</th>\n", |
| 735 |
" <th>coldspot</th>\n", |
| 736 |
" <th>peak</th>\n", |
| 737 |
" </tr>\n", |
| 738 |
" <tr>\n", |
| 739 |
" <th>Univariate</th>\n", |
| 740 |
" <th></th>\n", |
| 741 |
" <th></th>\n", |
| 742 |
" <th></th>\n", |
| 743 |
" <th></th>\n", |
| 744 |
" </tr>\n", |
| 745 |
" </thead>\n", |
| 746 |
" <tbody>\n", |
| 747 |
" <tr>\n", |
| 748 |
" <th>hotspot</th>\n", |
| 749 |
" <td>14</td>\n", |
| 750 |
" <td>3</td>\n", |
| 751 |
" <td>0</td>\n", |
| 752 |
" <td>0</td>\n", |
| 753 |
" </tr>\n", |
| 754 |
" <tr>\n", |
| 755 |
" <th>pit</th>\n", |
| 756 |
" <td>4</td>\n", |
| 757 |
" <td>7</td>\n", |
| 758 |
" <td>2</td>\n", |
| 759 |
" <td>0</td>\n", |
| 760 |
" </tr>\n", |
| 761 |
" <tr>\n", |
| 762 |
" <th>coldspot</th>\n", |
| 763 |
" <td>0</td>\n", |
| 764 |
" <td>5</td>\n", |
| 765 |
" <td>19</td>\n", |
| 766 |
" <td>3</td>\n", |
| 767 |
" </tr>\n", |
| 768 |
" <tr>\n", |
| 769 |
" <th>peak</th>\n", |
| 770 |
" <td>2</td>\n", |
| 771 |
" <td>0</td>\n", |
| 772 |
" <td>1</td>\n", |
| 773 |
" <td>6</td>\n", |
| 774 |
" </tr>\n", |
| 775 |
" </tbody>\n", |
| 776 |
"</table>\n", |
| 777 |
"</div>" |
| 778 |
], |
| 779 |
"text/plain": [ |
| 780 |
"Auxiliary hotspot pit coldspot peak\n", |
| 781 |
"Univariate \n", |
| 782 |
"hotspot 14 3 0 0\n", |
| 783 |
"pit 4 7 2 0\n", |
| 784 |
"coldspot 0 5 19 3\n", |
| 785 |
"peak 2 0 1 6" |
| 786 |
] |
| 787 |
}, |
| 788 |
"execution_count": 25, |
| 789 |
"metadata": {}, |
| 790 |
"output_type": "execute_result" |
| 791 |
} |
| 792 |
], |
| 793 |
"source": [ |
| 794 |
"#| echo: false\n", |
| 795 |
"aux_moves" |
| 796 |
] |
| 797 |
}, |
| 798 |
{ |
| 799 |
"cell_type": "markdown", |
| 800 |
"metadata": {}, |
| 801 |
"source": [ |
| 802 |
"From these, you can immediately see that the partial statistics can only move laterally. No observations moved from coldspot to anything other than \"peak\". That is, $x$ only affects our judgement about $y$, not $\\mathbf{W}y$. In the bottom, however, you can see that moves exist across any quadrant. Although it is theoretically possible, no observations move in between hotspot and coldspot or peak and pit.\n", |
| 803 |
"\n", |
| 804 |
"# Conclusion\n", |
| 805 |
"\n", |
| 806 |
"In sum, the `Partial_Moran_Local` and `Auxiliary_Moran_Local` are great ways to control for \"exogenous\" variables $x$ that you think might be driving the pattern of your outcome. In this case, if you want to remove *everything* due to $x$ first, you should use the `Auxiliary_Moran_Local`, which is equivalent to regressing $x$ out of $y$ and analyzing the residuals for structure. However, it is often the case that we want to account for $x$ *without* assigning all of the variation in $y$ to $x$ alone. In this case, use the `Partial_Moran_Local` to recover an estimate of the relationship between a site's $y$ value and its surrounding $y$ values while fixing $x$ to a constant value. \n", |
| 807 |
"\n", |
| 808 |
"# A bit more formal aside: Fritsch-Waugh-Lovell to the rescue\n", |
| 809 |
"\n", |
| 810 |
"For the econometricians in the room, there is an even better grounding than the graph-based argument from above to use the partial coefficient. If $\\mathbf{X}$ is a collection of exogenous control matrices, we generally want to estimate $\\rho$ and its partial products/local effects from the following Moran-form regression: \n", |
| 811 |
"\n", |
| 812 |
"$$ \\mathbf{W}y = \\mathbf{X}\\beta + \\rho y + \\epsilon $$\n", |
| 813 |
"\n", |
| 814 |
"How can we \"get rid of\" $\\mathbf{X}\\beta$ in our estimator for $\\rho$? Well, the [Frisch-Waugh-Lovell](https://en.wikipedia.org/wiki/Frisch–Waugh–Lovell_theorem) lets us filter the regression to get back to a \"regular\" Moran-form regression with just $\\mathbf{W}y$ and $y$. To do this, let $\\mathbf{P}=\\mathbf{I}-\\mathbf{X}'(\\mathbf{X}'\\mathbf{X})^{-1}\\mathbf{X}'$, and use FWL to re-state the regression above into a regression about the effect of $y$ on $\\mathbf{W}y$ that is independent of $x$:\n", |
| 815 |
"\n", |
| 816 |
"$$ \\mathbf{P}\\mathbf{W}y = I_{y|x} \\mathbf{P} y + \\epsilon $$\n", |
| 817 |
"\n", |
| 818 |
"If you calculate this out, we get our estimator above. For the auxiliary regression strategy, you can think of $e$ as \"what remains of $y$ that is independent of $x$. So, we would state $I_{x \\rightarrow y}$ for our auxiliary regression would seek to find the effect of \"what remains of $y$ that is independent of $x$\" on the lag of \"what remains of $y$ that is independent of $x$.\"\n", |
| 819 |
"\n", |
| 820 |
"$$ \\mathbf{W}\\mathbf{P}y = I_{x \\rightarrow y}\\mathbf{P}y + \\epsilon$$\n", |
| 821 |
"\n", |
| 822 |
"The two are not equivalent because $\\mathbf{WP}\\neq\\mathbf{PW}$ in general, even for symmetric $\\mathbf{W}$ and $\\mathbf{P}$; in practice, $\\mathbf{P}$ is always symmetric, but $\\mathbf{W}$ rarely is. Generally, we want $\\mathbf{P}\\mathbf{W}y$ as our outcome, not $\\mathbf{W}\\mathbf{P}y$. \n" |
| 823 |
] |
| 824 |
} |
| 825 |
], |
| 826 |
"metadata": { |
| 827 |
"kernelspec": { |
| 828 |
"display_name": "Analysis", |
| 829 |
"language": "python", |
| 830 |
"name": "analysis" |
| 831 |
}, |
| 832 |
"language_info": { |
| 833 |
"codemirror_mode": { |
| 834 |
"name": "ipython", |
| 835 |
"version": 3 |
| 836 |
}, |
| 837 |
"file_extension": ".py", |
| 838 |
"mimetype": "text/x-python", |
| 839 |
"name": "python", |
| 840 |
"nbconvert_exporter": "python", |
| 841 |
"pygments_lexer": "ipython3", |
| 842 |
"version": "3.11.11" |
| 843 |
} |
| 844 |
}, |
| 845 |
"nbformat": 4, |
| 846 |
"nbformat_minor": 4 |
| 847 |
} |