notebooks/03.07-Merge-and-Join.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<!--NAVIGATION-->\n",
    "< [Combining Datasets: Concat and Append](03.06-Concat-And-Append.ipynb) | [Contents](Index.ipynb) | [Aggregation and Grouping](03.08-Aggregation-and-Grouping.ipynb) >\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Combining Datasets: Merge and Join"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One essential feature offered by Pandas is its high-performance, in-memory join and merge operations.\n",
    "If you have ever worked with databases, you should be familiar with this type of data interaction.\n",
    "The main interface for this is the ``pd.merge`` function, and we'll see few examples of how this can work in practice.\n",
    "\n",
    "For convenience, we will start by redefining the ``display()`` functionality from the previous section:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "class display(object):\n",
    "    \"\"\"Display HTML representation of multiple objects\"\"\"\n",
    "    template = \"\"\"<div style=\"float: left; padding: 10px;\">\n",
    "    <p style='font-family:\"Courier New\", Courier, monospace'>{0}</p>{1}\n",
    "    </div>\"\"\"\n",
    "    def __init__(self, *args):\n",
    "        self.args = args\n",
    "        \n",
    "    def _repr_html_(self):\n",
    "        return '\\n'.join(self.template.format(a, eval(a)._repr_html_())\n",
    "                         for a in self.args)\n",
    "    \n",
    "    def __repr__(self):\n",
    "        return '\\n\\n'.join(a + '\\n' + repr(eval(a))\n",
    "                           for a in self.args)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Relational Algebra\n",
    "\n",
    "The behavior implemented in ``pd.merge()`` is a subset of what is known as *relational algebra*, which is a formal set of rules for manipulating relational data, and forms the conceptual foundation of operations available in most databases.\n",
    "The strength of the relational algebra approach is that it proposes several primitive operations, which become the building blocks of more complicated operations on any dataset.\n",
    "With this lexicon of fundamental operations implemented efficiently in a database or other program, a wide range of fairly complicated composite operations can be performed.\n",
    "\n",
    "Pandas implements several of these fundamental building-blocks in the ``pd.merge()`` function and the related ``join()`` method of ``Series`` and ``Dataframe``s.\n",
    "As we will see, these let you efficiently link data from different sources."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Categories of Joins\n",
    "\n",
    "The ``pd.merge()`` function implements a number of types of joins: the *one-to-one*, *many-to-one*, and *many-to-many* joins.\n",
    "All three types of joins are accessed via an identical call to the ``pd.merge()`` interface; the type of join performed depends on the form of the input data.\n",
    "Here we will show simple examples of the three types of merges, and discuss detailed options further below."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One-to-one joins\n",
    "\n",
    "Perhaps the simplest type of merge expresion is the one-to-one join, which is in many ways very similar to the column-wise concatenation seen in [Combining Datasets: Concat & Append](03.06-Concat-And-Append.ipynb).\n",
    "As a concrete example, consider the following two ``DataFrames`` which contain information on several employees in a company:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = pd.DataFrame({'employee': ['Bob', 'Jake', 'Lisa', 'Sue'],\n",
    "                    'group': ['Accounting', 'Engineering', 'Engineering', 'HR']})\n",
    "df2 = pd.DataFrame({'employee': ['Lisa', 'Bob', 'Jake', 'Sue'],\n",
    "                    'hire_date': [2004, 2008, 2012, 2014]})\n",
    "display('df1', 'df2')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To combine this information into a single ``DataFrame``, we can use the ``pd.merge()`` function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df3 = pd.merge(df1, df2)\n",
    "df3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The ``pd.merge()`` function recognizes that each ``DataFrame`` has an \"employee\" column, and automatically joins using this column as a key.\n",
    "The result of the merge is a new ``DataFrame`` that combines the information from the two inputs.\n",
    "Notice that the order of entries in each column is not necessarily maintained: in this case, the order of the \"employee\" column differs between ``df1`` and ``df2``, and the ``pd.merge()`` function correctly accounts for this.\n",
    "Additionally, keep in mind that the merge in general discards the index, except in the special case of merges by index (see the ``left_index`` and ``right_index`` keywords, discussed momentarily)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Many-to-one joins"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Many-to-one joins are joins in which one of the two key columns contains duplicate entries.\n",
    "For the many-to-one case, the resulting ``DataFrame`` will preserve those duplicate entries as appropriate.\n",
    "Consider the following example of a many-to-one join:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df4 = pd.DataFrame({'group': ['Accounting', 'Engineering', 'HR'],\n",
    "                    'supervisor': ['Carly', 'Guido', 'Steve']})\n",
    "display('df3', 'df4', 'pd.merge(df3, df4)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The resulting ``DataFrame`` has an aditional column with the \"supervisor\" information, where the information is repeated in one or more locations as required by the inputs."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Many-to-many joins"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Many-to-many joins are a bit confusing conceptually, but are nevertheless well defined.\n",
    "If the key column in both the left and right array contains duplicates, then the result is a many-to-many merge.\n",
    "This will be perhaps most clear with a concrete example.\n",
    "Consider the following, where we have a ``DataFrame`` showing one or more skills associated with a particular group.\n",
    "By performing a many-to-many join, we can recover the skills associated with any individual person:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df5 = pd.DataFrame({'group': ['Accounting', 'Accounting',\n",
    "                              'Engineering', 'Engineering', 'HR', 'HR'],\n",
    "                    'skills': ['math', 'spreadsheets', 'coding', 'linux',\n",
    "                               'spreadsheets', 'organization']})\n",
    "display('df1', 'df5', \"pd.merge(df1, df5)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These three types of joins can be used with other Pandas tools to implement a wide array of functionality.\n",
    "But in practice, datasets are rarely as clean as the one we're working with here.\n",
    "In the following section we'll consider some of the options provided by ``pd.merge()`` that enable you to tune how the join operations work."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Specification of the Merge Key"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We've already seen the default behavior of ``pd.merge()``: it looks for one or more matching column names between the two inputs, and uses this as the key.\n",
    "However, often the column names will not match so nicely, and ``pd.merge()`` provides a variety of options for handling this."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The ``on`` keyword\n",
    "\n",
    "Most simply, you can explicitly specify the name of the key column using the ``on`` keyword, which takes a column name or a list of column names:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df1', 'df2', \"pd.merge(df1, df2, on='employee')\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This option works only if both the left and right ``DataFrame``s have the specified column name."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The ``left_on`` and ``right_on`` keywords\n",
    "\n",
    "At times you may wish to merge two datasets with different column names; for example, we may have a dataset in which the employee name is labeled as \"name\" rather than \"employee\".\n",
    "In this case, we can use the ``left_on`` and ``right_on`` keywords to specify the two column names:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df3 = pd.DataFrame({'name': ['Bob', 'Jake', 'Lisa', 'Sue'],\n",
    "                    'salary': [70000, 80000, 120000, 90000]})\n",
    "display('df1', 'df3', 'pd.merge(df1, df3, left_on=\"employee\", right_on=\"name\")')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The result has a redundant column that we can drop if desired–for example, by using the ``drop()`` method of ``DataFrame``s:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.merge(df1, df3, left_on=\"employee\", right_on=\"name\").drop('name', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The ``left_index`` and ``right_index`` keywords\n",
    "\n",
    "Sometimes, rather than merging on a column, you would instead like to merge on an index.\n",
    "For example, your data might look like this:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1a = df1.set_index('employee')\n",
    "df2a = df2.set_index('employee')\n",
    "display('df1a', 'df2a')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can use the index as the key for merging by specifying the ``left_index`` and/or ``right_index`` flags in ``pd.merge()``:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df1a', 'df2a',\n",
    "        \"pd.merge(df1a, df2a, left_index=True, right_index=True)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For convenience, ``DataFrame``s implement the ``join()`` method, which performs a merge that defaults to joining on indices:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df1a', 'df2a', 'df1a.join(df2a)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you'd like to mix indices and columns, you can combine ``left_index`` with ``right_on`` or ``left_on`` with ``right_index`` to get the desired behavior:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df1a', 'df3', \"pd.merge(df1a, df3, left_index=True, right_on='name')\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All of these options also work with multiple indices and/or multiple columns; the interface for this behavior is very intuitive.\n",
    "For more information on this, see the [\"Merge, Join, and Concatenate\" section](http://pandas.pydata.org/pandas-docs/stable/merging.html) of the Pandas documentation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Specifying Set Arithmetic for Joins"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In all the preceding examples we have glossed over one important consideration in performing a join: the type of set arithmetic used in the join.\n",
    "This comes up when a value appears in one key column but not the other. Consider this example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df6 = pd.DataFrame({'name': ['Peter', 'Paul', 'Mary'],\n",
    "                    'food': ['fish', 'beans', 'bread']},\n",
    "                   columns=['name', 'food'])\n",
    "df7 = pd.DataFrame({'name': ['Mary', 'Joseph'],\n",
    "                    'drink': ['wine', 'beer']},\n",
    "                   columns=['name', 'drink'])\n",
    "display('df6', 'df7', 'pd.merge(df6, df7)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we have merged two datasets that have only a single \"name\" entry in common: Mary.\n",
    "By default, the result contains the *intersection* of the two sets of inputs; this is what is known as an *inner join*.\n",
    "We can specify this explicitly using the ``how`` keyword, which defaults to ``\"inner\"``:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.merge(df6, df7, how='inner')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Other options for the ``how`` keyword are ``'outer'``, ``'left'``, and ``'right'``.\n",
    "An *outer join* returns a join over the union of the input columns, and fills in all missing values with NAs:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df6', 'df7', \"pd.merge(df6, df7, how='outer')\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The *left join* and *right join* return joins over the left entries and right entries, respectively.\n",
    "For example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df6', 'df7', \"pd.merge(df6, df7, how='left')\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The output rows now correspond to the entries in the left input. Using\n",
    "``how='right'`` works in a similar manner.\n",
    "\n",
    "All of these options can be applied straightforwardly to any of the preceding join types."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Overlapping Column Names: The ``suffixes`` Keyword"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, you may end up in a case where your two input ``DataFrame``s have conflicting column names.\n",
    "Consider this example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df8 = pd.DataFrame({'name': ['Bob', 'Jake', 'Lisa', 'Sue'],\n",
    "                    'rank': [1, 2, 3, 4]})\n",
    "df9 = pd.DataFrame({'name': ['Bob', 'Jake', 'Lisa', 'Sue'],\n",
    "                    'rank': [3, 1, 4, 2]})\n",
    "display('df8', 'df9', 'pd.merge(df8, df9, on=\"name\")')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Because the output would have two conflicting column names, the merge function automatically appends a suffix ``_x`` or ``_y`` to make the output columns unique.\n",
    "If these defaults are inappropriate, it is possible to specify a custom suffix using the ``suffixes`` keyword:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "display('df8', 'df9', 'pd.merge(df8, df9, on=\"name\", suffixes=[\"_L\", \"_R\"])')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These suffixes work in any of the possible join patterns, and work also if there are multiple overlapping columns."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For more information on these patterns, see [Aggregation and Grouping](03.08-Aggregation-and-Grouping.ipynb) where we dive a bit deeper into relational algebra.\n",
    "Also see the [Pandas \"Merge, Join and Concatenate\" documentation](http://pandas.pydata.org/pandas-docs/stable/merging.html) for further discussion of these topics."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Example: US States Data\n",
    "\n",
    "Merge and join operations come up most often when combining data from different sources.\n",
    "Here we will consider an example of some data about US states and their populations.\n",
    "The data files can be found at http://github.com/jakevdp/data-USstates/:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Following are shell commands to download the data\n",
    "# !curl -O https://raw.githubusercontent.com/jakevdp/data-USstates/master/state-population.csv\n",
    "# !curl -O https://raw.githubusercontent.com/jakevdp/data-USstates/master/state-areas.csv\n",
    "# !curl -O https://raw.githubusercontent.com/jakevdp/data-USstates/master/state-abbrevs.csv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's take a look at the three datasets, using the Pandas ``read_csv()`` function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pop = pd.read_csv('data/state-population.csv')\n",
    "areas = pd.read_csv('data/state-areas.csv')\n",
    "abbrevs = pd.read_csv('data/state-abbrevs.csv')\n",
    "\n",
    "display('pop.head()', 'areas.head()', 'abbrevs.head()')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Given this information, say we want to compute a relatively straightforward result: rank US states and territories by their 2010 population density.\n",
    "We clearly have the data here to find this result, but we'll have to combine the datasets to find the result.\n",
    "\n",
    "We'll start with a many-to-one merge that will give us the full state name within the population ``DataFrame``.\n",
    "We want to merge based on the ``state/region``  column of ``pop``, and the ``abbreviation`` column of ``abbrevs``.\n",
    "We'll use ``how='outer'`` to make sure no data is thrown away due to mismatched labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged = pd.merge(pop, abbrevs, how='outer',\n",
    "                  left_on='state/region', right_on='abbreviation')\n",
    "merged = merged.drop('abbreviation', 1) # drop duplicate info\n",
    "merged.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's double-check whether there were any mismatches here, which we can do by looking for rows with nulls:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged.isnull().any()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some of the ``population`` info is null; let's figure out which these are!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged[merged['population'].isnull()].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It appears that all the null population values are from Puerto Rico prior to the year 2000; this is likely due to this data not being available from the original source.\n",
    "\n",
    "More importantly, we see also that some of the new ``state`` entries are also null, which means that there was no corresponding entry in the ``abbrevs`` key!\n",
    "Let's figure out which regions lack this match:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged.loc[merged['state'].isnull(), 'state/region'].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can quickly infer the issue: our population data includes entries for Puerto Rico (PR) and the United States as a whole (USA), while these entries do not appear in the state abbreviation key.\n",
    "We can fix these quickly by filling in appropriate entries:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged.loc[merged['state/region'] == 'PR', 'state'] = 'Puerto Rico'\n",
    "merged.loc[merged['state/region'] == 'USA', 'state'] = 'United States'\n",
    "merged.isnull().any()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No more nulls in the ``state`` column: we're all set!\n",
    "\n",
    "Now we can merge the result with the area data using a similar procedure.\n",
    "Examining our results, we will want to join on the ``state`` column in both:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "final = pd.merge(merged, areas, on='state', how='left')\n",
    "final.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Again, let's check for nulls to see if there were any mismatches:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "final.isnull().any()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are nulls in the ``area`` column; we can take a look to see which regions were ignored here:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "final['state'][final['area (sq. mi)'].isnull()].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that our ``areas`` ``DataFrame`` does not contain the area of the United States as a whole.\n",
    "We could insert the appropriate value (using the sum of all state areas, for instance), but in this case we'll just drop the null values because the population density of the entire United States is not relevant to our current discussion:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "final.dropna(inplace=True)\n",
    "final.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we have all the data we need. To answer the question of interest, let's first select the portion of the data corresponding with the year 2000, and the total population.\n",
    "We'll use the ``query()`` function to do this quickly (this requires the ``numexpr`` package to be installed; see [High-Performance Pandas: ``eval()`` and ``query()``](03.12-Performance-Eval-and-Query.ipynb)):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data2010 = final.query(\"year == 2010 & ages == 'total'\")\n",
    "data2010.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's compute the population density and display it in order.\n",
    "We'll start by re-indexing our data on the state, and then compute the result:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data2010.set_index('state', inplace=True)\n",
    "density = data2010['population'] / data2010['area (sq. mi)']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "density.sort_values(ascending=False, inplace=True)\n",
    "density.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The result is a ranking of US states plus Washington, DC, and Puerto Rico in order of their 2010 population density, in residents per square mile.\n",
    "We can see that by far the densest region in this dataset is Washington, DC (i.e., the District of Columbia); among states, the densest is New Jersey.\n",
    "\n",
    "We can also check the end of the list:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "density.tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that the least dense state, by far, is Alaska, averaging slightly over one resident per square mile.\n",
    "\n",
    "This type of messy data merging is a common task when trying to answer questions using real-world data sources.\n",
    "I hope that this example has given you an idea of the ways you can combine tools we've covered in order to gain insight from your data!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<!--NAVIGATION-->\n",
    "< [Combining Datasets: Concat and Append](03.06-Concat-And-Append.ipynb) | [Contents](Index.ipynb) | [Aggregation and Grouping](03.08-Aggregation-and-Grouping.ipynb) >\n",
    "\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}