Merge branch 'dataprod'

* dataprod: More yhat stuff Add annual mean pollution example Start playing with yhat
btq · Apr 25, 2014 · 0597b45 · 0597b45
2 parents 11d7470 + 51d8496
commit 0597b45
Show file tree

Hide file tree

Showing 8 changed files with 69,990 additions and 0 deletions.
diff --git a/09_DevelopingDataProducts/yhat/AP_example.R b/09_DevelopingDataProducts/yhat/AP_example.R
@@ -0,0 +1,99 @@
+## Create dataset of PM and O3 for all US taking year 2013 (annual
+## data from EPA)
+
+## This uses data from
+## http://aqsdr1.epa.gov/aqsweb/aqstmp/airdata/download_files.html
+
+## Read in the 2013 Annual Data
+
+d <- read.csv("annual_all_2013.csv", nrow = 68210)
+
+sub <- subset(d, Parameter.Name %in% c("PM2.5 - Local Conditions", "Ozone")
+              & Pullutant.Standard %in% c("Ozone 8-Hour 2008", "PM25 Annual 2006"),
+              c(State.Code, County.Code, Site.Num, Longitude, Latitude, Parameter.Name, Arithmetic.Mean, County.Name, City.Name, Pullutant.Standard))
+
+pollavg <- aggregate(sub[, "Arithmetic.Mean"],
+                     sub[, c("Longitude", "Latitude", "Parameter.Name")],
+                     mean, na.rm = TRUE)
+pollavg$Parameter.Name <- factor(pollavg$Parameter.Name, labels = c("ozone", "pm25"))
+names(pollavg)[4] <- "level"
+
+## Remove unneeded objects
+rm(d, sub)
+
+## library(fields)
+rdist <- function (x1, x2, miles = TRUE, R = NULL) {
+    if (is.null(R)) {
+        if (miles) 
+            R <- 3963.34
+        else R <- 6378.388
+    }
+    coslat1 <- cos((x1[, 2] * pi)/180)
+    sinlat1 <- sin((x1[, 2] * pi)/180)
+    coslon1 <- cos((x1[, 1] * pi)/180)
+    sinlon1 <- sin((x1[, 1] * pi)/180)
+    if (missing(x2)) {
+        pp <- cbind(coslat1 * coslon1, coslat1 * sinlon1, sinlat1) %*% 
+            t(cbind(coslat1 * coslon1, coslat1 * sinlon1, sinlat1))
+        return(R * acos(ifelse(abs(pp) > 1, 1 * sign(pp), pp)))
+    }
+    else {
+        coslat2 <- cos((x2[, 2] * pi)/180)
+        sinlat2 <- sin((x2[, 2] * pi)/180)
+        coslon2 <- cos((x2[, 1] * pi)/180)
+        sinlon2 <- sin((x2[, 1] * pi)/180)
+        pp <- cbind(coslat1 * coslon1, coslat1 * sinlon1, sinlat1) %*% 
+            t(cbind(coslat2 * coslon2, coslat2 * sinlon2, sinlat2))
+        return(R * acos(ifelse(abs(pp) > 1, 1 * sign(pp), pp)))
+    }
+}
+
+monitors <- data.matrix(pollavg[, c("Longitude", "Latitude")])
+
+
+pollutant <- function(df) {
+        x <- data.matrix(df[, c("lon", "lat")])
+        r <- df$radius
+        d <- rdist(monitors, x)
+        use <- lapply(seq_len(ncol(d)), function(i) {
+                which(d[, i] < r[i])
+        })
+        levels <- sapply(use, function(idx) {
+                with(pollavg[idx, ], tapply(level, Parameter.Name, mean))
+        })
+        dlevel <- as.data.frame(t(levels))
+        data.frame(df, dlevel)
+}
+
+library(yhatr)
+
+model.transform <- function(df) {
+        df
+}
+
+model.predict <- function(df) {
+        pollutant(df)
+}
+
+yhat.config  <- c(
+        username="[email protected]",
+        apikey="90d2a80bb532cabb2387aa51ac4553cc",
+        env="http://sandbox.yhathq.com/"
+)
+
+yhat.deploy("pollutant")
+
+
+
+################################################################################
+## Client side
+
+library(yhatr)
+yhat.config  <- c(
+        username="[email protected]",
+        apikey="90d2a80bb532cabb2387aa51ac4553cc",
+        env="http://sandbox.yhathq.com/"
+)
+df <- data.frame(lon = c(-76.6167, -118.25), lat = c(39.2833, 34.05),
+                 radius = 20)
+yhat.predict("pollutant", df)
diff --git a/09_DevelopingDataProducts/yhat/annual_all_2013.csv b/09_DevelopingDataProducts/yhat/annual_all_2013.csv
diff --git a/09_DevelopingDataProducts/yhat/environment.R b/09_DevelopingDataProducts/yhat/environment.R
@@ -0,0 +1,22 @@
+library(yhatr)
+
+model.require <- function() {
+
+}
+
+model.transform <- function(df) {
+        df
+}
+
+model.predict <- function(df) {
+        data.frame(version = R.version.string, stringsAsFactors = FALSE)
+}
+
+yhat.config  <- c(
+        username="[email protected]",
+        apikey="90d2a80bb532cabb2387aa51ac4553cc",
+        env="http://sandbox.yhathq.com/"
+)
+
+
+yhat.deploy("environment")
diff --git a/09_DevelopingDataProducts/yhat/front.html b/09_DevelopingDataProducts/yhat/front.html
@@ -0,0 +1,30 @@
+<!DOCTYPE HTML>
+<html>
+    <head>
+        <meta name="viewport" content="width=device-width">
+        <title>What Am I Breathing?</title>
+        <link href="/static/css/main.css" rel="stylesheet" type="text/css"></link>
+    </head>
+    <body>
+        <div class="container">
+            <form id="rec">
+	      Latitude <input id="lat" type="text" /><br />
+	      Longitude <input id="lon" type="text" /><br />
+	      Radius (mi) <input id="radius" type="text" /><br />
+              <button class="btn" type="submit">Go!</button>
+            </form>
+            <section class="results">
+                <p class="lead">Top Ten Recommended Beers</p>
+                <figure class="visualization row">
+                    <div class="col-sm-6 col-1"></div>
+                    <div class="col-sm-6 col-2"></div>
+                </figure>
+            </section>
+            <pre id="result" class="language-javascript"><code class="language-javascript"></code></pre>
+        </div>
+        <script src="/static/js/jquery-1.10.2.min.js"></script>
+        <script src="/static/js/code.js"></script>
+        <script src="/static/js/beers.js"></script>
+        <script src="/static/js/main.js"></script>
+    </body>
+</html>
diff --git a/09_DevelopingDataProducts/yhat/ozone.R b/09_DevelopingDataProducts/yhat/ozone.R
@@ -0,0 +1,24 @@
+library(yhatr)
+
+model.require <- function() {
+
+}
+
+model.transform <- function(df) {
+        df
+}
+
+model.predict <- function(df) {
+        data.frame(Ozone = predict(fit, newdata = df))
+}
+
+fit <- lm(Ozone ~ Wind + Temp, data = airquality)
+
+yhat.config  <- c(
+        username="[email protected]",
+        apikey="90d2a80bb532cabb2387aa51ac4553cc",
+        env="http://sandbox.yhathq.com/"
+)
+
+
+yhat.deploy("ozone")
diff --git a/09_DevelopingDataProducts/yhat/pollutant.sh b/09_DevelopingDataProducts/yhat/pollutant.sh
@@ -0,0 +1,4 @@
+#!/bin/bash
+
+curl -X POST -H "Content-Type: application/json" --user [email protected]:90d2a80bb532cabb2387aa51ac4553cc --data '{"lon": $1, "lat": $2, "radius": $3 }' http://sandbox.yhathq.com/[email protected]/models/pollutant/
+
diff --git a/09_DevelopingDataProducts/yhat/predictozone.sh b/09_DevelopingDataProducts/yhat/predictozone.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+
+echo $1
+echo $2
+
+curl -X POST -H "Content-Type: application/json" \
+    --user [email protected]:90d2a80bb532cabb2387aa51ac4553cc \
+    --data '{"Wind": $1, "Temp": $2 }' \
+    http://sandbox.yhathq.com/[email protected]/models/ozone/
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,4 @@
		#!/bin/bash

		curl -X POST -H "Content-Type: application/json" --user [email protected]:90d2a80bb532cabb2387aa51ac4553cc --data '{"lon": $1, "lat": $2, "radius": $3 }' http://sandbox.yhathq.com/[email protected]/models/pollutant/