diff --git a/docs/examples/img/predictive_elicitation/ppa_preliz.png b/docs/examples/img/predictive_elicitation/ppa_preliz.png
new file mode 100644
index 00000000..b9c56992
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/ppa_preliz.png differ
diff --git a/docs/examples/img/predictive_elicitation/predictive_explorer_bambi_custplot.png b/docs/examples/img/predictive_elicitation/predictive_explorer_bambi_custplot.png
new file mode 100644
index 00000000..be7a9610
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/predictive_explorer_bambi_custplot.png differ
diff --git a/docs/examples/img/predictive_elicitation/predictive_explorer_bambi_ecdf.png b/docs/examples/img/predictive_elicitation/predictive_explorer_bambi_ecdf.png
new file mode 100644
index 00000000..a2985d42
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/predictive_explorer_bambi_ecdf.png differ
diff --git a/docs/examples/img/predictive_elicitation/predictive_explorer_preliz_custplot.png b/docs/examples/img/predictive_elicitation/predictive_explorer_preliz_custplot.png
new file mode 100644
index 00000000..5ecc8bb4
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/predictive_explorer_preliz_custplot.png differ
diff --git a/docs/examples/img/predictive_elicitation/predictive_explorer_preliz_ecdf.png b/docs/examples/img/predictive_elicitation/predictive_explorer_preliz_ecdf.png
new file mode 100644
index 00000000..55319ee7
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/predictive_explorer_preliz_ecdf.png differ
diff --git a/docs/examples/img/predictive_elicitation/predictive_explorer_pymc_custplot.png b/docs/examples/img/predictive_elicitation/predictive_explorer_pymc_custplot.png
new file mode 100644
index 00000000..84f3b5fb
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/predictive_explorer_pymc_custplot.png differ
diff --git a/docs/examples/img/predictive_elicitation/predictive_explorer_pymc_ecdf.png b/docs/examples/img/predictive_elicitation/predictive_explorer_pymc_ecdf.png
new file mode 100644
index 00000000..23b0b5b1
Binary files /dev/null and b/docs/examples/img/predictive_elicitation/predictive_explorer_pymc_ecdf.png differ
diff --git a/docs/examples/observed_space_examples.ipynb b/docs/examples/observed_space_examples.ipynb
deleted file mode 100644
index d7ca9e08..00000000
--- a/docs/examples/observed_space_examples.ipynb
+++ /dev/null
@@ -1,2497 +0,0 @@
-{
- "cells": [
- {
- "cell_type": "markdown",
- "id": "82281aab",
- "metadata": {},
- "source": [
- "# Predictive elicitation"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "f06af29a",
- "metadata": {},
- "source": [
- "PyMC and Bambi are optional dependencies of PreliZ, you only need them if you want to use the `predictive_explorer` function with PyMC or Bambi models.\n",
- "\n",
- "To install them, you can run the following command:\n",
- "\n",
- "```bash\n",
- "conda install -c conda-forge pymc bambi\n",
- "```"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 1,
- "id": "6b551879",
- "metadata": {},
- "outputs": [],
- "source": [
- "import arviz as az\n",
- "import preliz as pz\n",
- "import numpy as np\n",
- "import pandas as pd\n",
- "\n",
- "# Optional dependencies\n",
- "import pymc as pm\n",
- "import bambi as bmb"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 2,
- "id": "9875c4c5",
- "metadata": {},
- "outputs": [],
- "source": [
- "az.style.library[\"arviz-doc\"][\"figure.dpi\"] = 100\n",
- "az.style.library[\"arviz-doc\"][\"figure.figsize\"] = (10, 4)\n",
- "az.style.use(\"arviz-doc\")"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "baca7bd9",
- "metadata": {},
- "source": [
- "Predictive elicitation refers to the elicitation of priors by making evaluations on the observed space, instead of the parameters space. The advantage of the former is that generally it is easier to think about observations than parameters, specially for domain experts. If you are a forest engineer, you may be more familiar with the expected rate of wildfires in an area than with the `rho` parameter of a model.\n",
- "\n",
- "Currently PreliZ provides 3 functions for predictive elicitation. Most of the future development of PreliZ will be focused on improving these tools and providing other tools for predictive elicitation."
- ]
- },
- {
- "cell_type": "markdown",
- "id": "74dadce9",
- "metadata": {},
- "source": [
- "(predictive_explorer)=\n",
- "\n",
- "## Predictive explorer\n",
- "\n",
- "This function will automatically generate textboxes for a function, which makes it easier to explore how the prior predictive distribution changes when we change the priors.\n",
- "\n",
- "So suppose you already have a model, but you are unsure about the implications of its parameters. You can write it using PreliZ distributions and put it inside a function with the arguments being the parameters you want to explore."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 3,
- "id": "7f275a20",
- "metadata": {},
- "outputs": [],
- "source": [
- "x = np.linspace(0, 1, 100)\n",
- "\n",
- "def a_preliz_model(a_mu, a_sigma, c_sigma=1):\n",
- " a = pz.Normal(a_mu, a_sigma).rvs()\n",
- " c = pz.Gamma(mu=2, sigma=c_sigma).rvs()\n",
- " b = pz.Normal(np.exp(a)*x, c).rvs()\n",
- " return b"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "dd8232ac",
- "metadata": {},
- "source": [
- "By calling `predictive_explorer` you will get textboxes with some default initial values and for you to explore."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 4,
- "id": "6cce4214",
- "metadata": {
- "scrolled": true
- },
- "outputs": [
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "1ef4c6b0ae144464bc517880013e0b80",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(VBox(children=(FloatText(value=0.0, description='a_mu (-inf, inf)', step=0.1, style=Description…"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
- "source": [
- "pz.predictive_explorer(a_preliz_model)"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "009ab7a8",
- "metadata": {},
- "source": [
- "You can also write the model using PyMC distributions and place it inside the function with the arguments being the parameters you want to explore."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 5,
- "id": "c6913d8a",
- "metadata": {},
- "outputs": [],
- "source": [
- "x = np.linspace(0, 1, 100)\n",
- "\n",
- "def a_pymc_model(a_mu, a_sigma, c_sigma=1):\n",
- " with pm.Model() as model:\n",
- " a = pm.Normal(\"a\", a_mu, a_sigma)\n",
- " c = pm.Gamma(\"c\", mu=2, sigma=c_sigma)\n",
- " b = pm.Normal(\"b\", np.exp(a) * x, c, observed=[0] * 100)\n",
- " return model"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "d933d854",
- "metadata": {},
- "source": [
- "The `predictive_explorer` function auto-detects that the model contains PyMC distributions, alternatively you can specify that the function should use the PyMC engine by providing the parameter `engine=pymc`."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 6,
- "id": "05475814",
- "metadata": {
- "scrolled": true
- },
- "outputs": [
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "3b3b4902a9454942bc40d51182ccc00a",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(VBox(children=(FloatText(value=0.0, description='a_mu (-inf, inf)', step=0.1, style=Description…"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
- "source": [
- "pz.predictive_explorer(a_pymc_model)"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "d4609392",
- "metadata": {},
- "source": [
- "Similarly, you can also write the model using Bambi Distributions."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 7,
- "id": "cb6eb44d",
- "metadata": {},
- "outputs": [],
- "source": [
- "data = pd.DataFrame(\n",
- " {\n",
- " \"y\": np.random.normal(size=100),\n",
- " \"x\": np.random.normal(size=100),\n",
- " }\n",
- ")\n",
- "def a_bambi_model(a_mu, a_sigma):\n",
- " prior = {\"Intercept\": bmb.Prior(\"Normal\", mu=a_mu, sigma=a_sigma)}\n",
- " a_model = bmb.Model(\"y ~ x\", data, priors=prior)\n",
- " return a_model"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "f47ccd4d",
- "metadata": {},
- "source": [
- "Similar to PyMC, the `predictive_explorer` function auto-detects that the model contains Bambi distributions, alternatively you can specify that the function should use the Bambi engine by providing the parameter `engine=bambi`."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 8,
- "id": "cbd5ccfa",
- "metadata": {},
- "outputs": [
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "4458c19484644bd7bafb09344d6f34d0",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(VBox(children=(FloatText(value=0.0, description='a_mu (-inf, inf)', step=0.1, style=Description…"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
- "source": [
- "pz.predictive_explorer(a_bambi_model)"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "3c0e068d-e833-462f-83ea-d681f3dcea71",
- "metadata": {},
- "source": [
- " After the parameter name, you will see a tuple indicating the valid range of the parameter. These values are inferred automatically by `predictive_explorer`. Currently, we use a very simple heuristic to find the range, so take the suggestion with a pinch of salt"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "e6ccc87a-673b-4cc0-9af9-372d05b9fb07",
- "metadata": {},
- "source": [
- "`predictive_explorer` supports three types of plots: empirical cumulative distribution functions (CDFs), kernel density estimations (KDEs), and histograms. \n",
- "Additionally, you can also add custom Matplotlib code using the `plot_func` parameter."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 9,
- "id": "482e8511-fcf4-4562-988d-8aebb8e3dfda",
- "metadata": {},
- "outputs": [],
- "source": [
- "# Custom function to plot a histogram\n",
- "def custom_hist(predictions, ax):\n",
- " ax.hist(predictions.flatten(), bins='auto', alpha=0.7)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 10,
- "id": "dfec9aa1-0585-4f96-a24c-f29165e0a71a",
- "metadata": {},
- "outputs": [
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "76d0e113fb33439489f1427e190ae13d",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(VBox(children=(FloatText(value=0.0, description='a_mu (-inf, inf)', step=0.1, style=Description…"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
- "source": [
- "# plot_func set to custom_hist for preliz model\n",
- "pz.predictive_explorer(a_preliz_model, samples=10, plot_func=custom_hist)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 11,
- "id": "05617d8b",
- "metadata": {},
- "outputs": [
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "c14b49e294db4e1397c222081b7ee4af",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(VBox(children=(FloatText(value=0.0, description='a_mu (-inf, inf)', step=0.1, style=Description…"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
- "source": [
- "# plot_func set to custom_hist for pymc model\n",
- "pz.predictive_explorer(a_pymc_model, samples=10, plot_func=custom_hist)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 12,
- "id": "dfb027dc",
- "metadata": {},
- "outputs": [
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "5e7b5cd8a6aa4eab9c63286653377b9f",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(VBox(children=(FloatText(value=0.0, description='a_mu (-inf, inf)', step=0.1, style=Description…"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- }
- ],
- "source": [
- "# plot_func set to custom_hist for bambi model\n",
- "pz.predictive_explorer(a_bambi_model, samples=10, plot_func=custom_hist)"
- ]
- },
- {
- "cell_type": "markdown",
- "id": "ce4f1868",
- "metadata": {},
- "source": [
- "(prior_predictive_assistant)=\n",
- "\n",
- "## Prior predictive assistant\n",
- "\n",
- "Now we will see a very experimental tool called \"prior predictive assistant\" (ppa for short). Suppose we have a model and we want to explore the prior predictive distribution and if possible refine the priors in our model. A common scenario in practice is that after some thinking we are able to define a model, but our priors are still too wide, as we are underusing our prior knowledge.\n",
- "\n",
- "Let assume a Normal model with an unknown mean and standard deviation, here `a` and `b` represent our priors and `y` the likelihood."
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 13,
- "id": "c138324f",
- "metadata": {},
- "outputs": [],
- "source": [
- "np.random.seed(1237)\n",
- "def a_preliz_model():\n",
- " a = pz.Normal(0, 10).rvs()\n",
- " b = pz.HalfNormal(10).rvs()\n",
- " y = pz.Normal(a, b).rvs(100)\n",
- " return a, b, y"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 14,
- "id": "9d1d7547",
- "metadata": {},
- "outputs": [
- {
- "name": "stderr",
- "output_type": "stream",
- "text": [
- "\"This is an experimental method under development, use with caution.\n"
- ]
- },
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "25946ab210c34a2fb6a72b948db5556a",
- "version_major": 2,
- "version_minor": 0
- },
- "text/plain": [
- "HBox(children=(Text(value='0', description='references: ', layout=Layout(margin='0 20px 0 0', width='230px'), …"
- ]
- },
- "metadata": {},
- "output_type": "display_data"
- },
- {
- "data": {
- "application/vnd.jupyter.widget-view+json": {
- "model_id": "1fca9bf602264901a9296a287cc3c7de",
- "version_major": 2,
- "version_minor": 0
- },
- "image/png": "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",
- "text/html": [
- "\n",
- " \n",
- "
\n",
- " Figure\n",
- "
\n",
- "
\n",
- "
\n",
- "
\n",
- " Figure\n",
- "
\n",
- "
\n",
- "