Merge branch 'master' of https://github.com/stan-dev/example-models

Author: mitzimorris

jupyter/covid-inf-rate/SantaClara_CmdStanPy.ipynb

@@ -6,7 +6,7 @@
       "name": "SantaClara_CmdStanPy.ipynb",
       "provenance": [],
       "collapsed_sections": [],
-      "authorship_tag": "ABX9TyNh1NuXuIIVzc5g9FQpoq1m",
+      "authorship_tag": "ABX9TyN1oagX6vCOD9nEfpF/t/vS",
       "include_colab_link": true
     },
     "kernelspec": {
@@ -373,6 +373,16 @@
       "execution_count": 0,
       "outputs": []
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "L1CzZED7FtCS",
+        "colab_type": "text"
+      },
+      "source": [
+        "In order to avoid divergent iterations, which result in biased estimates, we set the parameter adapt_delta to 0.98.  To check the resulting sample, we run CmdStan's `diagnose` utility."
+      ]
+    },
     {
       "cell_type": "code",
       "metadata": {
@@ -502,13 +512,85 @@
       "execution_count": 0,
       "outputs": []
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "vGBOK0c4ekFv",
+        "colab_type": "text"
+      },
+      "source": [
+        "### Play around with the models and data\n",
+        "\n",
+        "If you want to play around with your models and data, you can do so using Colab's `file.upload` utility."
+      ]
+    },
     {
       "cell_type": "code",
       "metadata": {
         "id": "nZh7IS4fl8VM",
         "colab_type": "code",
         "colab": {}
       },
+      "source": [
+        "from google.colab import files"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "V38jwvTvfbtQ",
+        "colab_type": "text"
+      },
+      "source": [
+        "The `files.upload()` function allows you to choose files on your machine to upload. It returns a Python dictionary which is a map of filenames to their contents."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "QCjxWPoKfVoE",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "uploaded = files.upload()\n",
+        "uploaded"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "50vnyQN3gObz",
+        "colab_type": "text"
+      },
+      "source": [
+        "The uploaded files are also saved into the current working directory, so you can refer to them directly by filename. Therefore, once you've uploaded a new Stan program file, you can use this colab notebook to fit your model to the data."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "POhsju4ffld2",
+        "colab_type": "code",
+        "colab": {}
+      },
+      "source": [
+        "!ls -lt | head -3"
+      ],
+      "execution_count": 0,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "w5zVoY_8gTF7",
+        "colab_type": "code",
+        "colab": {}
+      },
       "source": [
         ""
       ],

jupyter/covid-inf-rate/SantaClara_CmdStanR.ipynb

@@ -350,6 +350,16 @@
       "execution_count": 0,
       "outputs": []
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "yWNEf0sdC1Kk",
+        "colab_type": "text"
+      },
+      "source": [
+        "In order to avoid divergent iterations, which result in biased estimates, we set the parameter `adapt_delta` to 0.98."
+      ]
+    },
     {
       "cell_type": "code",
       "metadata": {
@@ -359,7 +369,7 @@
       },
       "source": [
         "hier_model <- cmdstan_model(stan_file='hier_sens_spec.stan')\n",
-        "hier_fit <- hier_model$sample(data='santa_clara_all.data.json')\n"
+        "hier_fit <- hier_model$sample(data='santa_clara_all.data.json', adapt_delta=0.98)\n"
       ],
       "execution_count": 0,
       "outputs": []
@@ -421,7 +431,7 @@
       },
       "source": [
         "hier_model_v2 <- cmdstan_model(stan_file='hier_sens_spec_offset_mult.stan')\n",
-        "hier_fit_v2 <- hier_model_v2$sample(data='santa_clara_all.data.json')"
+        "hier_fit_v2 <- hier_model_v2$sample(data='santa_clara_all.data.json', adapt_delta=0.98)"
       ],
       "execution_count": 0,
       "outputs": []