{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Data Analysis using Jupyter Notebooks Part 2\n",
"\n",
"Benjamin J. Morgan\n",
"\n",
"# Tutorial 3"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 2D numpy arrays\n",
"\n",
"In all the examples using `numpy` arrays that you have seen so far, these arrays have been **one dimensional**.\n",
"\n",
"It can often be useful to collect different data together in a table. \n",
"One way to do this is by combining one-dimensional `numpy` arrays into larger, **two-dimensional**, arrays.\n",
"\n",
">```python\n",
"import numpy as np\n",
"# create an array `x` with the integers 1 to 5\n",
"x = np.arange(1,6)\n",
"# create three new arrays by performing calculations on `x`\n",
"y1 = x**2\n",
"y2 = x+3\n",
"y3 = x/2 + 1\n",
"print('x=',x)\n",
"print('y1=',y1)\n",
"print('y2=',y2)\n",
"print('y3=',y3)\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"You can combine numpy arrays into a table as **columns** using `np.column_stack()`\n",
"\n",
">```python\n",
"# combine x, y1, y2, and y3 as columns in a new table\n",
"column_table = np.column_stack( ( x, y1, y2, y3 ) )\n",
"print( column_table )\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Or as **rows** using np.row_stack()\n",
"\n",
">```python\n",
"# arrange x, y1, y2, and y3 as rows in a new table\n",
"row_table = np.row_stack( ( x, y1, y2, y3 ) )\n",
"print( row_table )\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Remember that a 1D `numpy` array can be indexed like a list.\n",
">```python\n",
"my_1D_array = np.array( [ 1, 2, 3, 4, 5, 6] )\n",
"my_1D_array[2:5] \n",
"# [2:5] selects from 2 jumps, up to, but not including, 5 jumps\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"A 2D `numpy` array can be treated like a list of lists, and indexing returns selected rows.\n",
">```python\n",
"row_table[1] # return the 2nd row (1 jump from the start)\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Because each row is a 1D `numpy` array, you can use a second index to select a single entry.\n",
">```python\n",
"row_table[1][3]\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"These two indices can be combined into a single bracket, where the first value selects one of more rows, and the second value selects one or more columns.\n",
">```python\n",
"row_table[1,3]\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"To select a single row, we make use of the range character `:`. Remember, for a list or 1D array, `:` lets us select a range of elements, and leaving out one of the numbers selects all elements up to the start, or end, of the list.\n",
"\n",
">```python\n",
"my_list = [ 'a', 'b', 'c', 'd', 'e' ]\n",
"my_list[1:]\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Leaving out *both* numbers extends our selection up to both ends of the list or array.\n",
">```python\n",
"my_list[:]\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For a 2D array, you can think of this as “every row” or “every column”.\n",
"\n",
">```python\n",
"print( row_table )\n",
"print()\n",
"print( row_table[:,3] ) # all rows, jump 3 columns\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"Use a combination of row and column indexing to select [ 6., 7., 8.] from row_table\n",
"
"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"You can also convert between 1D and 2D numpy arrays (or even higher dimensions) using `np.reshape()`\n",
"\n",
">```python\n",
"a = ( [ 1, 2, 3, 4, 5, 6 ] )\n",
"print( a )\n",
"b = np.reshape( a, (3, 2) ) # reshape into a 3×2 2D array\n",
"print( b )\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"and back again\n",
"\n",
">```python\n",
"c = np.reshape( b, (6) )\n",
"print( c )\n",
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"Create a numpy 1D array containing the numbers 1–100.
\n",
"Reshape this 1D array into a 10×10 2D array. \n",
"
"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.1"
}
},
"nbformat": 4,
"nbformat_minor": 2
}