notes-2024-07-16.ipynb

notes.2023-07-12.ipynb

notes.2023-08-07.ipynb

notes.2024-07-16.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "604c1eec",
   "metadata": {},
   "source": [
    "# How do I even get started with Data Visualization?\n",
    "\n",
    "## Agenda\n",
    "\n",
    "0. Why bother with programmatic visualization at all?\n",
    "1. The types of visualizations\n",
    "2. The types of data-viz tools\n",
    "3. Demo & categorize data-viz tools\n",
    "4. Your turn!\n",
    "5. Final discusison & demonstrations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8d2a131a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import display\n",
    "\n",
    "display(\"Let's Get Started\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab7f49a3",
   "metadata": {},
   "source": [
    "## Google Colab: Install Missing Packages\n",
    "\n",
    "Google Colab comes with many Python packages pre-installed, but we will need the\n",
    "following pacakges as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8c1d2eb4",
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install flexitext==0.2.0"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67445e46",
   "metadata": {},
   "source": [
    "## Types of Visualizations\n",
    "- Exploratory\n",
    "- Communicative\n",
    "\n",
    "## Why even use programming in the first place?\n",
    "\n",
    "## Types of Data Visualization Tools\n",
    "- Non-programmatic\n",
    "    - low : hand/computer drawing\n",
    "    - high: GUI chart builders* (excel, tableau, ...)\n",
    "- Programmatic\n",
    "    - low : programmatic drawing\n",
    "    - high: declarative & convenience"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eca0d185",
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import display\n",
    "from pandas import read_csv\n",
    "\n",
    "anscombe = read_csv('https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/anscombe.csv')\n",
    "\n",
    "display(\n",
    "    # anscombe.head(),\n",
    "    # anscombe.tail(),\n",
    "    anscombe.groupby('id')[['x', 'y']].agg(['mean', 'std']),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10a39e18",
   "metadata": {},
   "source": [
    "## Common Data Visualization APIs\n",
    "\n",
    "### Drawing (non-declarative)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "07bd9d3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas import read_csv\n",
    "from matplotlib.pyplot import subplots, show\n",
    "\n",
    "anscombe = read_csv('https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/anscombe.csv')\n",
    "\n",
    "# print(anscombe['id'].unique())\n",
    "fig, axes = subplots(nrows=2, ncols=2)\n",
    "\n",
    "for (label, group), ax in zip(anscombe.groupby('id'), axes.flat):\n",
    "    ax.scatter(group['x'], group['y'], s=24)\n",
    "    ax.set_title(label)\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d2d22ced",
   "metadata": {},
   "source": [
    "### High Level - Declarative"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a83af00d",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pyplot import show\n",
    "from plotnine import ggplot, facet_wrap, geom_point, geom_smooth, labs, theme_minimal, aes\n",
    "from pandas import read_csv\n",
    "\n",
    "anscombe = read_csv('https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/anscombe.csv')\n",
    "\n",
    "fig = (\n",
    "    ggplot(anscombe, aes(x='x', y='y'))\n",
    "    + facet_wrap('id', ncol=2)\n",
    "    + geom_point()\n",
    "    + labs(x='x variable', y='y variable', title='Anscombe’s Quartet')\n",
    "    + theme_minimal()\n",
    ")\n",
    "\n",
    "print(fig)\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8abd78d1",
   "metadata": {},
   "source": [
    "### High Level - Convenience"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0462567b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pyplot import show\n",
    "from seaborn import lmplot\n",
    "from pandas import read_csv\n",
    "\n",
    "anscombe = read_csv('https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/anscombe.csv')\n",
    "\n",
    "fg = lmplot(anscombe, x='x', y='y', col='id', col_wrap=2)\n",
    "print(fg)\n",
    "# show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d7c1239f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from seaborn.objects import Plot, Dots, PolyFit, Line\n",
    "from pandas import read_csv\n",
    "\n",
    "anscombe = read_csv('https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/anscombe.csv')\n",
    "\n",
    "(\n",
    "    Plot(anscombe, x='x', y='y')\n",
    "    .facet(col='id', wrap=2)\n",
    "    .add(Dots(color='black'))\n",
    "    .add(Line(), PolyFit(1))\n",
    ").show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c28501cb",
   "metadata": {},
   "source": [
    "### Is it worth it to learn multiple data visualization libraries/languages?\n",
    "\n",
    "## Let’s Make Some Viz!\n",
    "\n",
    "### Static Data Visualizations (explore, communicative, fun)\n",
    "\n",
    "**matplotlib**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a0080d56",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pyplot import figure, show\n",
    "from matplotlib.patches import Circle, Rectangle\n",
    "\n",
    "fig = figure(figsize=(6,6))\n",
    "\n",
    "c = Circle((.5, .8), .1)\n",
    "fig.add_artist(c)\n",
    "\n",
    "## uncomment below if running from Jupyter Notebook/Google Colab\n",
    "ax = fig.add_axes([0, 0, 0, 0])\n",
    "ax.set_visible(False)\n",
    "\n",
    "# body_rect = Rectangle((.47, .75), .06, -.5)\n",
    "# fig.add_artist(body_rect)\n",
    "\n",
    "# arms_rect = Rectangle((.3, .55), .4, .05)\n",
    "# fig.add_artist(arms_rect)\n",
    "\n",
    "# lleg_rect = Rectangle((.5, .3), .3, .05, angle=225)\n",
    "# fig.add_artist(lleg_rect)\n",
    "\n",
    "# rleg_rect = Rectangle((.47, .26), .3, .05, angle=-45)\n",
    "# fig.add_artist(rleg_rect)\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17b7e400",
   "metadata": {},
   "source": [
    "**Useful Things to draw for Data Viz**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f40441fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "from numpy import linspace, pi, sin, cos\n",
    "from matplotlib.pyplot import figure, show, plot, rc\n",
    "\n",
    "rc('font', size=16)\n",
    "\n",
    "xs = linspace(0, 2 * pi)\n",
    "\n",
    "fig = figure()\n",
    "                   #X   Y   W   H\n",
    "ax = fig.add_axes([.3, .3, .5, .5])\n",
    "\n",
    "ax.plot(xs, sin(xs))\n",
    "ax.plot(xs, cos(xs))\n",
    "\n",
    "ax.set_ylabel('this is my y label')\n",
    "fig.supylabel('this my figure y label')\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c0e14c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import display\n",
    "\n",
    "from numpy import linspace, pi, sin, cos\n",
    "from matplotlib.pyplot import subplots, show\n",
    "\n",
    "# fig, ax = subplots()\n",
    "\n",
    "# fig, axes = subplots(nrows=1, ncols=2)\n",
    "fig, axes = subplots(nrows=2, ncols=2)\n",
    "# print(axes)\n",
    "# print(axes[:, 0])\n",
    "\n",
    "xs = linspace(0, 2 * pi)\n",
    "\n",
    "# display(axes)\n",
    "display(type(axes))\n",
    "\n",
    "for ax in axes[0, :]:\n",
    "    ax.plot(xs, sin(xs))\n",
    "\n",
    "\n",
    "# display(axes, type(axes), axes[0])\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec54e83b",
   "metadata": {},
   "source": [
    "Matplotlib is object oriented\n",
    "- Containers → Artists\n",
    "- Figure → Axes →\n",
    "    - X/YAxis\n",
    "        - ticks\n",
    "        - ticklabels\n",
    "        - axis label\n",
    "    - Primitives\n",
    "        - Patches (Circle, Rectangle)\n",
    "        - Line2d\n",
    "        - Annotations/Text\n",
    "        - ...\n",
    "    - Legend\n",
    "        - Primitives\n",
    "        - Text (label & title)\n",
    "\n",
    "- Coordinate Spaces: values → ... → screen\n",
    "    - Proportional coordinate space (Figure & Axes)\n",
    "    - Data coordinate space         (Axes)\n",
    "    - Identity/point space          (Figure)\n",
    "\n",
    "**Applied to Star Trader Data - Tracking Ship Failures**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "639b0358",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pyplot import subplots, show\n",
    "from pandas import read_csv, to_datetime\n",
    "\n",
    "df = (\n",
    "    read_csv(\n",
    "        'https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/failures.csv',\n",
    "        index_col=['date', 'player','ship'],\n",
    "        parse_dates=['date'],\n",
    "    )\n",
    "    .sort_index()\n",
    ")\n",
    "\n",
    "plot_data = (\n",
    "    df.pivot_table(index='date', columns='player', values='faults', aggfunc='sum')\n",
    "    .rolling('90D').mean()\n",
    ")\n",
    "\n",
    "ax = plot_data.plot(legend=False)\n",
    "for line in ax.lines:\n",
    "    x, y = line.get_data()\n",
    "    ax.annotate(\n",
    "        line.get_label(), xy=(x[-1], y[-1]),\n",
    "        xytext=(5, 0), textcoords='offset points',\n",
    "        color=line.get_color()\n",
    "    )\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee66b5fd",
   "metadata": {},
   "source": [
    "### Interactive Data Visualizations (explore, fun)\n",
    "\n",
    "- System/function Exploration\n",
    "- Data Exploration\n",
    "- System Observability\n",
    "\n",
    "*Limited Communicatve Ability unless STRONGLY guided*\n",
    "\n",
    "**bokeh & panel** - a powerful way to share your data on the web!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "04a3f3af",
   "metadata": {},
   "outputs": [],
   "source": [
    "from panel import extension\n",
    "\n",
    "# Increase font size of widgets\n",
    "css = '''\n",
    ".bk-root .bk, .bk-root .bk:before, .bk-root .bk:after {\n",
    "  font-size: 110%;\n",
    "}\n",
    "'''\n",
    "extension(raw_css=[css]) # Connect `panel` application to notebook runtime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b35b4016",
   "metadata": {},
   "outputs": [],
   "source": [
    "from panel import Column\n",
    "from bokeh.plotting import figure\n",
    "from bokeh.models import ColumnDataSource\n",
    "\n",
    "from numpy import linspace, zeros\n",
    "from scipy.stats import skewnorm\n",
    "\n",
    "loc   = 5\n",
    "scale = 1\n",
    "skew  = 0\n",
    "\n",
    "cds = ColumnDataSource({\n",
    "    'x': linspace(-10, 10, 500),\n",
    "    'y1': zeros(shape=500),\n",
    "})\n",
    "cds.data['y2'] = skewnorm(loc=loc, scale=scale, a=skew).pdf(cds.data['x'])\n",
    "\n",
    "def update_plot(loc, scale, skew):\n",
    "    cds.data['y2'] = skewnorm.pdf(x=cds.data['x'], a=skew, loc=loc, scale=scale)\n",
    "\n",
    "p = figure(y_range=(0, .5), width=500, height=300)\n",
    "p.varea(x='x', y1='y1', y2='y2', source=cds, alpha=.3)\n",
    "p.line(x='x', y='y2', source=cds, line_width=4)\n",
    "p.yaxis.major_label_text_font_size = \"20pt\"\n",
    "p.xaxis.major_label_text_font_size = \"20pt\"\n",
    "\n",
    "Column(p).servable()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "119af6e3",
   "metadata": {},
   "source": [
    "adding interactivity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "320bab7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from panel import Column, bind\n",
    "from panel.widgets import FloatSlider\n",
    "from bokeh.plotting import figure\n",
    "from bokeh.models import ColumnDataSource\n",
    "\n",
    "from numpy import linspace, zeros\n",
    "from scipy.stats import skewnorm\n",
    "\n",
    "loc   = FloatSlider(name='mean', value=0, start=-10, end=10)\n",
    "scale = FloatSlider(name='std. dev', value=1, start=.1, end=10)\n",
    "skew  = FloatSlider(name='skew', value=0, start=-6, end=6)\n",
    "\n",
    "# Data abstraction\n",
    "cds = ColumnDataSource({\n",
    "    'x': linspace(-10, 10, 500),\n",
    "    'y1': zeros(shape=500),\n",
    "    'y2': zeros(shape=500),\n",
    "})\n",
    "\n",
    "def update_plot(loc, scale, skew):\n",
    "    cds.data['y2'] = skewnorm.pdf(x=cds.data['x'], a=skew, loc=loc, scale=scale)\n",
    "\n",
    "p = figure(y_range=(0, .5), width=500, height=300)\n",
    "p.varea(x='x', y1='y1', y2='y2', source=cds, alpha=.3)\n",
    "p.line(x='x', y='y2', source=cds, line_width=4)\n",
    "p.yaxis.major_label_text_font_size = \"20pt\"\n",
    "p.xaxis.major_label_text_font_size = \"20pt\"\n",
    "\n",
    "Column(\n",
    "    Column(loc, scale, skew), # render widgets\n",
    "    p,                        # render plot\n",
    "    bind(update_plot, skew=skew, loc=loc, scale=scale) # bind widgets to `update_plot` function\n",
    ").servable()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68868054",
   "metadata": {},
   "source": [
    "**Applied to our Star Trader Data - Planetary Weather**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8d5425bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas import read_csv\n",
    "\n",
    "df = (\n",
    "    read_csv(\n",
    "        'https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/weather_york.csv',\n",
    "        usecols=['date', 'temperature_max', 'temperature_min'],\n",
    "        parse_dates=['date'],\n",
    "        index_col='date'\n",
    "    )\n",
    ").loc['1990':'2000']\n",
    "\n",
    "# Long timeseries Zoom\n",
    "from bokeh.models import RangeTool\n",
    "from bokeh.plotting import figure, ColumnDataSource\n",
    "from panel import Column\n",
    "from pandas import to_datetime, DateOffset\n",
    "\n",
    "cds = ColumnDataSource(df)\n",
    "p = figure(\n",
    "    width=1000, height=500, x_axis_type='datetime', y_range=[0, 110],\n",
    "    x_range=[df.index.min(), df.index.min() + DateOffset(years=1, days=-1)],\n",
    ")\n",
    "p.vbar(x='date', bottom='temperature_min', top='temperature_max', source=cds, width=(24 * 60 * 60 * 1000))\n",
    "\n",
    "range_p = figure(\n",
    "    width=p.width, height=p.height // 2, x_axis_type='datetime', y_range=[0, 110],\n",
    "    x_range=[df.index.min(), df.index.max()],\n",
    ")\n",
    "range_p.vbar(x='date', bottom='temperature_min', top='temperature_max', source=cds, width=(24 * 60 * 60 * 1000))\n",
    "\n",
    "rangetool = RangeTool(x_range=p.x_range)\n",
    "range_p.add_tools(rangetool)\n",
    "\n",
    "Column(p, range_p).servable()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee6be309",
   "metadata": {},
   "source": [
    "**Looking for more resources to help get you started?** - try recreating examples from the documentation for these tools\n",
    "(see **Useful Links** below), or follow along with a tutorial for a tool of your choice!\n",
    "\n",
    "## Useful Links\n",
    "\n",
    "### Conceptual Guides\n",
    "\n",
    "[Data to Viz](https://www.data-to-viz.com/) provides a flowchart-style to the\n",
    "types of charts one can create given various types of data.\n",
    "\n",
    "[R Graph Gallery](https://r-graph-gallery.com/) and its counter part [Python Graph Gallery](https://python-graph-gallery.com/) provide a vast number of high quality charts written in either R or Python.\n",
    "\n",
    "### Tools\n",
    "\n",
    "**Matplotlib**\n",
    "- Tutorial: https://matplotlib.org/stable/tutorials/index.html\n",
    "- Cheatsheets: https://matplotlib.org/cheatsheets/\n",
    "- Examples: https://matplotlib.org/stable/gallery/index.html\n",
    "\n",
    "**Plotnine**\n",
    "- Tutorial: http://r-statistics.co/Complete-Ggplot2-Tutorial-Part1-With-R-Code.html (note that plotnine does not have official tutorials, so please refer to ggplot2)\n",
    "- Examples: https://plotnine.readthedocs.io/en/stable/gallery.html#\n",
    "\n",
    "**Bokeh**\n",
    "- Tutorial: https://docs.bokeh.org/en/latest/docs/first_steps.html#first-steps\n",
    "- Examples: https://docs.bokeh.org/en/latest/docs/gallery.html#gallery\n",
    "\n",
    "**Seaborn**\n",
    "- Tutorial: https://seaborn.pydata.org/tutorial/introduction.html\n",
    "- Examples: https://seaborn.pydata.org/examples/index.html\n",
    "\n",
    "## Your Turn…\n",
    "\n",
    "Take any of the tools we have discussed today and make 1 static or 1 interactive (web) chart.\n",
    "Remember before you start, think about what you want to create? Something exploratory, communicative?\n",
    "\n",
    "### Suggested Starting Points\n",
    "\n",
    "*static* using data/weather_york.csv datasets\n",
    "- Explore: visualize as much of the data as possible, do you notice any trends?\n",
    "- Communicate: Select one interesting feature to highlight and create a chart\n",
    "  that communiates that feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7aa0b19a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from warnings import simplefilter\n",
    "simplefilter('ignore')\n",
    "from pandas import read_csv\n",
    "\n",
    "df = (\n",
    "    read_csv(\n",
    "        'https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/weather_york.csv',\n",
    "        usecols=['date', 'temperature_max', 'temperature_min'],\n",
    "        parse_dates=['date'],\n",
    "    )\n",
    ")\n",
    "\n",
    "from matplotlib.pyplot import subplots, show\n",
    "from pandas import DateOffset\n",
    "\n",
    "interest = 2022\n",
    "\n",
    "fig, ax = subplots()\n",
    "for year, group in df.groupby(df['date'].dt.year):\n",
    "    normed_year = group['date'] + DateOffset(year=1900)\n",
    "    if year == interest:\n",
    "        ax.plot(normed_year, group['temperature_max'], alpha=1, lw=1, color='tab:red')\n",
    "    else:\n",
    "        ax.plot(normed_year, group['temperature_max'], alpha=.1, lw=.5, color='gray')\n",
    "\n",
    "from matplotlib.dates import MonthLocator, DateFormatter\n",
    "from matplotlib.ticker import NullFormatter\n",
    "\n",
    "ax.xaxis.set_major_locator(MonthLocator())\n",
    "ax.xaxis.set_major_formatter(NullFormatter())\n",
    "ax.xaxis.set_minor_locator(MonthLocator(bymonthday=15))\n",
    "ax.xaxis.set_minor_formatter(DateFormatter('%b'))\n",
    "ax.xaxis.set_tick_params(which='both', length=0)\n",
    "\n",
    "ax.margins(x=0)\n",
    "\n",
    "from itertools import pairwise\n",
    "for i, (left, right) in enumerate(pairwise(ax.get_xticks())):\n",
    "    if i % 2 == 0:\n",
    "        ax.axvspan(left, right, 0, 1, color='gainsboro', alpha=.2)\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9ae3180",
   "metadata": {},
   "outputs": [],
   "source": [
    "from warnings import simplefilter\n",
    "simplefilter('ignore')\n",
    "from pandas import read_csv\n",
    "\n",
    "df = (\n",
    "    read_csv(\n",
    "        'https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/weather_york.csv',\n",
    "        usecols=['date', 'temperature_max', 'temperature_min'],\n",
    "        parse_dates=['date'],\n",
    "    )\n",
    "    .assign(\n",
    "        year= lambda d: d['date'].dt.year,\n",
    "        month=lambda d: d['date'].dt.month,\n",
    "        day=  lambda d: d['date'].dt.day,\n",
    "    )\n",
    ")\n",
    "\n",
    "ref_years = (1960, 1990)\n",
    "show_year = 2002\n",
    "\n",
    "ref_raw = df.loc[lambda d: d['year'].between(*ref_years)]\n",
    "show_raw = df.loc[lambda d: d['year'] == show_year]\n",
    "\n",
    "plot_df = (\n",
    "    ref_raw.groupby(['month', 'day'])['temperature_max'].agg(\n",
    "        ref_ub  =lambda g: g.quantile(.95),\n",
    "        ref_lb  =lambda g: g.quantile(.05),\n",
    "        ref_mean='mean',\n",
    "    )\n",
    "    .merge(show_raw, left_index=True, right_on=['month', 'day'])\n",
    "    .assign(\n",
    "        distance=lambda d: d['temperature_max'] - d['ref_mean'],\n",
    "        norm_distance=lambda d: d['distance'] / d['distance'].abs().max()\n",
    "    )\n",
    ")\n",
    "\n",
    "from matplotlib.pyplot import subplots, show\n",
    "from matplotlib.colors import TwoSlopeNorm\n",
    "\n",
    "fig, ax = subplots(gridspec_kw={'top': .8, 'bottom': .2})\n",
    "lb_line, = ax.plot('date', 'ref_lb', color='k', data=plot_df, ls='--', lw=.7, zorder=6)\n",
    "ub_line, = ax.plot('date', 'ref_ub', color='k', data=plot_df, ls='--', lw=.7, zorder=6)\n",
    "context_pc = ax.fill_between(\n",
    "    'date', 'ref_lb', 'ref_ub', data=plot_df, color='gainsboro', ec='none', zorder=5\n",
    ")\n",
    "avg_line = ax.plot('date', 'ref_mean', data=plot_df, color='k', lw=1, zorder=6)\n",
    "\n",
    "raw_pc = ax.fill_between(\n",
    "    'date', 'temperature_max', 'ref_mean', data=plot_df, fc='none', ec=ax.get_facecolor(), zorder=5, lw=.1\n",
    ")\n",
    "\n",
    "arr = raw_pc.get_paths()[0].vertices\n",
    "(x0, y0), (x1, y1) = arr.min(axis=0), arr.max(axis=0)\n",
    "\n",
    "gradient = ax.imshow(\n",
    "    plot_df['norm_distance'].to_numpy().reshape(1, -1),\n",
    "    extent=[x0, x1, y0, y1],\n",
    "    aspect='auto',\n",
    "    cmap='RdBu_r',\n",
    "    norm=TwoSlopeNorm(0),\n",
    "    interpolation='bicubic',\n",
    "    zorder=5,\n",
    ")\n",
    "\n",
    "gradient.set_clip_path(raw_pc.get_paths()[0], transform=ax.transData)\n",
    "\n",
    "from matplotlib.dates import MonthLocator, DateFormatter\n",
    "from matplotlib.ticker import NullFormatter\n",
    "\n",
    "ax.xaxis.set_major_locator(MonthLocator())\n",
    "ax.xaxis.set_major_formatter(NullFormatter())\n",
    "ax.xaxis.set_minor_locator(MonthLocator(bymonthday=15))\n",
    "ax.xaxis.set_minor_formatter(DateFormatter('%b'))\n",
    "ax.xaxis.set_tick_params(which='both', length=0)\n",
    "\n",
    "ax.margins(x=0)\n",
    "\n",
    "from itertools import pairwise\n",
    "for i, (left, right) in enumerate(pairwise(ax.get_xticks())):\n",
    "    if i % 2 == 0:\n",
    "        ax.axvspan(left, right, 0, 1, color='gainsboro', alpha=.2)\n",
    "\n",
    "\n",
    "from flexitext import flexitext\n",
    "\n",
    "flexitext(\n",
    "    s=(\n",
    "        f'<size:medium,weight:bold>{show_year}</> <color:tab:red>Hot</> and <color:tab:blue>Cold</>'\n",
    "        ' Temperature deviations from historical average in planet York'\n",
    "    ),\n",
    "    x=0, y=1.01,\n",
    "    va='bottom',\n",
    "    ax=ax\n",
    ")\n",
    "\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ed04a51",
   "metadata": {},
   "source": [
    "*interactive* Using the failures.csv dataset,\n",
    "  - Plot the total failures for each 'player' for each day.\n",
    "    - Apply a smoothing factor (rolling average) of 90 days prior to plotting the data.\n",
    "  - Create a slider widget that control the number of days involved in the smoothing.\n",
    "      - e.g. this slider should allow me to apply 0 days of smoothing all the way up to 90 days of smoothing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "15b1a115",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pyplot import subplots, show\n",
    "from pandas import read_csv, to_datetime\n",
    "\n",
    "df = (\n",
    "    read_csv(\n",
    "        'https://raw.githubusercontent.com/dutc-io/agu-data-viz/main/data/failures.csv',\n",
    "        parse_dates=['date'],\n",
    "    )\n",
    ")\n",
    "\n",
    "print(df.head())"
   ]
  }
 ],
 "metadata": {
  "jupytext": {
   "cell_metadata_filter": "-all",
   "main_language": "python",
   "notebook_metadata_filter": "-all"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}

notes.2024-08-01.ipynb