lessons/06_Data_Processing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Data Processing\n",
    "## Computational Methods in Psychology and Neuroscience\n",
    "### Psychology 4215/7215 --- Fall 2023\n",
    "By: Per B. Sederberg, PhD\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Lesson Objectives\n",
    "\n",
    "Upon completion of this lesson, students should have learned:\n",
    "\n",
    "1. Read data from slog files\n",
    "2. The Series and DataFrame data structures in Pandas\n",
    "3. Load slogs as a DataFrame\n",
    "4. Some basic operations on the data\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Updating SMILE\n",
    "\n",
    "- First you can test whether there is a new version Kivy, which is the primary dependency of SMILE:\n",
    "\n",
    "```bash\n",
    "conda install -c conda-forge kivy\n",
    "```\n",
    "\n",
    "- Then you can update SMILE right from the GitHub repository (note the upgrade option at the end):\n",
    "\n",
    "```bash\n",
    "pip install git+https://github.com/compmem/smile --upgrade\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Math Distract\n",
    "\n",
    "- Sometimes you want to have a delay period, e.g., between study and test\n",
    "- Although it may be fine to have an empty delay, often you'd like to fill it with a task that prevents rehearsal of the studied items\n",
    "- We provide a subroutine that generates math problems!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from smile.common import *\n",
    "from smile.math_distract import MathDistract\n",
    "\n",
    "exp = Experiment(show_splash=False, resolution=(1024,768))\n",
    "\n",
    "Wait(1.0)\n",
    "MathDistract(num_vars=3,\n",
    "             min_num=1,\n",
    "             max_num=9,\n",
    "             max_probs=50,\n",
    "             duration=20)\n",
    "    \n",
    "exp.run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Moving Dot stimuli\n",
    "\n",
    "- A class of stimuli at the core of many studies of perceptual decision-making\n",
    "- We provide a custom state for it:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from smile.common import *\n",
    "from smile.moving_dots import MovingDots\n",
    "\n",
    "exp = Experiment(show_splash=False, resolution=(1024,768))\n",
    "\n",
    "# set up some config\n",
    "dot_speed = 180\n",
    "\n",
    "# set initial values\n",
    "exp.cr=0.2\n",
    "exp.cl=0.2\n",
    "motion_props = [{\"coherence\": exp.cr, \"direction\": 0, \"direction_variance\": 0},\n",
    "                {\"coherence\": exp.cl, \"direction\": 180, \"direction_variance\": 0}]\n",
    "with Loop():\n",
    "    with Parallel():\n",
    "        dots = MovingDots(color='white', scale=3, num_dots=100, radius=200,\n",
    "                                motion_props=motion_props, speed=dot_speed,\n",
    "                                lifespan=0.5, lifespan_variance=1.5)\n",
    "        lr = Label(text=Ref.object('Right Coherence:\\n{:.2f}').format(exp.cr), \n",
    "                   left=dots.right+40, font_size=28) \n",
    "        ll = Label(text=Ref.object('Left Coherence:\\n{:.2f}').format(exp.cl), \n",
    "                   right=dots.left-40, font_size=28)\n",
    "\n",
    "    with UntilDone():\n",
    "            kp = KeyPress(keys=['UP','DOWN','LEFT','RIGHT'])\n",
    "            with If(kp.pressed=='UP'):\n",
    "                exp.cr=exp.cr+0.05\n",
    "                exp.cl=exp.cl+0.05\n",
    "                with If(exp.cr+exp.cl>1.0):\n",
    "                    exp.cr=exp.cr-0.05\n",
    "                    exp.cl=exp.cl-0.05\n",
    "            with Elif((kp.pressed=='DOWN')):\n",
    "                exp.cr=exp.cr-0.05\n",
    "                exp.cl=exp.cl-0.05\n",
    "                with If(exp.cr<0.05):\n",
    "                    exp.cr=0.0\n",
    "                with If(exp.cl<0.05):\n",
    "                    exp.cl=0.0\n",
    "            with Elif(kp.pressed=='LEFT'):\n",
    "                exp.cl=exp.cl+0.05\n",
    "                with If(exp.cr+exp.cl>1.0):\n",
    "                    exp.cl=exp.cl-0.05\n",
    "            with Elif(kp.pressed=='RIGHT'):\n",
    "                exp.cr=exp.cr+0.05\n",
    "                with If(exp.cr+exp.cl>1.0):\n",
    "                    exp.cr=exp.cr-0.05\n",
    "            # update the motion props\n",
    "            dots.update(motion_props=[{\"coherence\": exp.cr, \"direction\": 0},\n",
    "                                      {\"coherence\": exp.cl, \"direction\": 180}])\n",
    "            lr.update(text=Ref.object('Right Coherence:\\n{:.2f}').format(exp.cr))\n",
    "            ll.update(text=Ref.object('Left Coherence:\\n{:.2f}').format(exp.cl))\n",
    "with UntilDone():\n",
    "    KeyPress(keys=['ENTER'])\n",
    "\n",
    "    \n",
    "exp.run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Reading in slog files\n",
    "\n",
    "- SMILE stores data in log files with the `.slog` file extension\n",
    "- It is a custom format that are pickled and compressed dictionaries\n",
    "- We can read them in with a SMILE function `log2dl` that converts the log to a list of dictionaries (i.e., a dict list):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'rt': 2.941615221919392,\n",
       "  'disappear_time_error': 0.0,\n",
       "  'disappear_time_time': 617.657840390051,\n",
       "  'run_num': 0,\n",
       "  'refresh_rate': 26.731467774154293,\n",
       "  'appear_time_error': 0.0,\n",
       "  'appear_time_time': 614.7072808193332,\n",
       "  'eeg_pulse_time': None,\n",
       "  'right_coherence': 0.06,\n",
       "  'left_coherence': 0.0,\n",
       "  'press_time_error': 0.0005104763318968253,\n",
       "  'press_time_time': 617.6488960412526,\n",
       "  'incorrect_resp': '1',\n",
       "  'pressed': '1',\n",
       "  'log_time': 618.1488960412526,\n",
       "  'correct_resp': '4',\n",
       "  'correct': False,\n",
       "  'fmri_tr_time': None,\n",
       "  'log_num': 0},\n",
       " {'rt': 1.5730700115865375,\n",
       "  'disappear_time_error': 0.0,\n",
       "  'disappear_time_time': 620.2500236883183,\n",
       "  'run_num': 0,\n",
       "  'refresh_rate': 26.59035375261749,\n",
       "  'appear_time_error': 0.0,\n",
       "  'appear_time_time': 618.6663967214259,\n",
       "  'eeg_pulse_time': None,\n",
       "  'right_coherence': 0.24,\n",
       "  'left_coherence': 0.3,\n",
       "  'press_time_error': 0.00047487459045214564,\n",
       "  'press_time_time': 620.2394667330125,\n",
       "  'incorrect_resp': '4',\n",
       "  'pressed': '4',\n",
       "  'log_time': 620.7394667330125,\n",
       "  'correct_resp': '1',\n",
       "  'correct': False,\n",
       "  'fmri_tr_time': None,\n",
       "  'log_num': 0}]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from smile.log import log2dl\n",
    "dl = log2dl('log_MD_0.slog')\n",
    "dl[:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Pandas\n",
    "\n",
    "- Library at the core of most data science with Python\n",
    "- Provides two key data structures: `Series` and `DataFrame`\n",
    "- The key feature of Pandas is that ***data alignment is intrinsic***. \n",
    "  - The link between labels and data will not be broken unless done so explicitly by you.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Series\n",
    "\n",
    "- A `Series` is a one-dimensional labeled array capable of holding any data type:\n",
    "  - integers, strings, floating point numbers, Python objects, etc...\n",
    "- The axis labels are collectively referred to as the index. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.065871\n",
       "1    2.134110\n",
       "2   -1.492454\n",
       "3   -0.201191\n",
       "4   -0.572541\n",
       "dtype: float64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "s = pd.Series(np.random.randn(5))\n",
    "s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    1.146421\n",
       "b   -0.309947\n",
       "c   -0.862864\n",
       "d    1.096532\n",
       "e   -1.431422\n",
       "dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# you can specify the index\n",
    "s = pd.Series(np.random.randn(5), index=['a', 'b', 'c', 'd', 'e'])\n",
    "s"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Series are ndarray-like\n",
    "\n",
    "- You can slice a series and it will also slice your index\n",
    "- And many of the same methods are available (e.g., mean, sum, etc...)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "c   -0.862864\n",
       "d    1.096532\n",
       "e   -1.431422\n",
       "dtype: float64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s[2:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.07225605972307125"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    1.146421\n",
       "d    1.096532\n",
       "dtype: float64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s[s > s.mean()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a     True\n",
       "b    False\n",
       "c    False\n",
       "d     True\n",
       "e    False\n",
       "dtype: bool"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s > s.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Series is also dict-like\n",
    "\n",
    "- A Series is like a fixed-size dict in that you can get and set values by index label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.096532092339073"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s['d']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'d' in s"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Series keeps array operations aligned\n",
    "\n",
    "- Series can also be passed into most NumPy methods expecting an ndarray.\n",
    "- Alignment will be maintained"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    2.292842\n",
       "b   -0.619895\n",
       "c   -1.725728\n",
       "d    2.193064\n",
       "e   -2.862845\n",
       "dtype: float64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s+s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.07225605972307125"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    3.146910\n",
       "b    0.733486\n",
       "c    0.421952\n",
       "d    2.993766\n",
       "e    0.238969\n",
       "dtype: float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a         NaN\n",
       "b   -0.619895\n",
       "c   -1.725728\n",
       "d    2.193064\n",
       "e         NaN\n",
       "dtype: float64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s[1:] + s[:-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     0\n",
       "1   -16\n",
       "2    16\n",
       "3     3\n",
       "4    16\n",
       "dtype: int64"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = pd.Series(np.random.randn(5)*10)\n",
    "x.astype(np.int)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## DataFrame\n",
    "\n",
    "- `DataFrame` is a 2-dimensional labeled data structure with columns of potentially different types. \n",
    "- You can think of it like a spreadsheet or SQL table, or a dict of `Series` objects.\n",
    "- It's possible to create a DataFrame a lot of different ways."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two\n",
       "a  1.0  4.0\n",
       "b  2.0  3.0\n",
       "c  3.0  2.0\n",
       "d  4.0  1.0"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# e.g., from a dictionary\n",
    "d = {'one': [1., 2., 3., 4.],\n",
    "     'two': [4., 3., 2., 1.]}\n",
    "df = pd.DataFrame(d, index=['a', 'b', 'c', 'd'])\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Column selection, addition, deletion\n",
    "\n",
    "- You can treat a DataFrame like a dict of Series objects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    1.0\n",
       "b    2.0\n",
       "c    3.0\n",
       "d    4.0\n",
       "Name: one, dtype: float64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pick a column\n",
    "df['one']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh\n",
       "a  1.0  4.0    5.0   False\n",
       "b  2.0  3.0    5.0   False\n",
       "c  3.0  2.0    5.0    True\n",
       "d  4.0  1.0    5.0    True"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# make new columns\n",
    "df['three'] = df['one'] + df['two']\n",
    "df['thresh'] = df['one'] > 2.0\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "one       float64\n",
       "two       float64\n",
       "three     float64\n",
       "thresh       bool\n",
       "dtype: object"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# you can make copies with the copy method\n",
    "df2 = df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>thresh</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  thresh\n",
       "a  1.0   False\n",
       "b  2.0   False\n",
       "c  3.0    True\n",
       "d  4.0    True"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# you can delete with del or pop\n",
    "del df['two']\n",
    "df.pop('three')\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh\n",
       "a  1.0  4.0    5.0   False\n",
       "b  2.0  3.0    5.0   False\n",
       "c  3.0  2.0    5.0    True\n",
       "d  4.0  1.0    5.0    True"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# df2 should be untouched\n",
    "df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# copy it back to retrieve the deleted columns\n",
    "df = df2.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "      <th>foo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh  foo\n",
       "a  1.0  4.0    5.0   False  bar\n",
       "b  2.0  3.0    5.0   False  bar\n",
       "c  3.0  2.0    5.0    True  bar\n",
       "d  4.0  1.0    5.0    True  bar"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# new values will populate the entire column\n",
    "df['foo'] = 'bar'\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Indexing and Selection\n",
    "\n",
    "  - loc\n",
    "  - iloc\n",
    "  - boolean\n",
    "  - column\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "      <th>foo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh  foo\n",
       "c  3.0  2.0    5.0    True  bar\n",
       "d  4.0  1.0    5.0    True  bar"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[df['one']>2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "      <th>foo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh  foo\n",
       "a  1.0  4.0    5.0   False  bar\n",
       "b  2.0  3.0    5.0   False  bar\n",
       "c  3.0  2.0    5.0    True  bar\n",
       "d  4.0  1.0    5.0    True  bar"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>thresh</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  thresh\n",
       "a  1.0   False\n",
       "b  2.0   False\n",
       "c  3.0    True\n",
       "d  4.0    True"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pick columns of interest\n",
    "df[['one', 'thresh']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "      <th>foo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh  foo\n",
       "b  2.0  3.0    5.0   False  bar\n",
       "c  3.0  2.0    5.0    True  bar"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pick rows by index\n",
    "df.loc[['b', 'c']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "      <th>foo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>True</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh  foo\n",
       "b  2.0  3.0    5.0   False  bar\n",
       "d  4.0  1.0    5.0    True  bar"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pick rows by numeric index\n",
    "df.iloc[[1, 3]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>thresh</th>\n",
       "      <th>foo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>False</td>\n",
       "      <td>bar</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three  thresh  foo\n",
       "a  1.0  4.0    5.0   False  bar"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pick rows by boolean index\n",
    "df2 = df[(df['two']>2) & (df['one']<=1)]\n",
    "df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   two  three\n",
       "a  4.0    5.0"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pick rows (and specific columns) by boolean index\n",
    "df3 = df.loc[(df['two']>2) & (df['one']<=1), ['two', 'three']]\n",
    "df3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## SMILE and Pandas\n",
    "\n",
    "- We can create a DataFrame from a dict list in SMILE:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rt</th>\n",
       "      <th>disappear_time_error</th>\n",
       "      <th>disappear_time_time</th>\n",
       "      <th>run_num</th>\n",
       "      <th>refresh_rate</th>\n",
       "      <th>appear_time_error</th>\n",
       "      <th>appear_time_time</th>\n",
       "      <th>eeg_pulse_time</th>\n",
       "      <th>right_coherence</th>\n",
       "      <th>left_coherence</th>\n",
       "      <th>press_time_error</th>\n",
       "      <th>press_time_time</th>\n",
       "      <th>incorrect_resp</th>\n",
       "      <th>pressed</th>\n",
       "      <th>log_time</th>\n",
       "      <th>correct_resp</th>\n",
       "      <th>correct</th>\n",
       "      <th>fmri_tr_time</th>\n",
       "      <th>log_num</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.941615</td>\n",
       "      <td>0.0</td>\n",
       "      <td>617.657840</td>\n",
       "      <td>0</td>\n",
       "      <td>26.731468</td>\n",
       "      <td>0.0</td>\n",
       "      <td>614.707281</td>\n",
       "      <td>None</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000510</td>\n",
       "      <td>617.648896</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>618.148896</td>\n",
       "      <td>4</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.573070</td>\n",
       "      <td>0.0</td>\n",
       "      <td>620.250024</td>\n",
       "      <td>0</td>\n",
       "      <td>26.590354</td>\n",
       "      <td>0.0</td>\n",
       "      <td>618.666397</td>\n",
       "      <td>None</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.000475</td>\n",
       "      <td>620.239467</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>620.739467</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.810989</td>\n",
       "      <td>0.0</td>\n",
       "      <td>624.100785</td>\n",
       "      <td>0</td>\n",
       "      <td>27.489738</td>\n",
       "      <td>0.0</td>\n",
       "      <td>621.275211</td>\n",
       "      <td>None</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000544</td>\n",
       "      <td>624.086200</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>624.586200</td>\n",
       "      <td>4</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.872496</td>\n",
       "      <td>0.0</td>\n",
       "      <td>625.751111</td>\n",
       "      <td>0</td>\n",
       "      <td>27.255089</td>\n",
       "      <td>0.0</td>\n",
       "      <td>624.850929</td>\n",
       "      <td>None</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.001483</td>\n",
       "      <td>625.723425</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>626.223425</td>\n",
       "      <td>1</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.818831</td>\n",
       "      <td>0.0</td>\n",
       "      <td>628.618369</td>\n",
       "      <td>0</td>\n",
       "      <td>26.489366</td>\n",
       "      <td>0.0</td>\n",
       "      <td>626.784654</td>\n",
       "      <td>None</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.000755</td>\n",
       "      <td>628.603485</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>629.103485</td>\n",
       "      <td>1</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         rt  disappear_time_error  disappear_time_time  run_num  refresh_rate  \\\n",
       "0  2.941615                   0.0           617.657840        0     26.731468   \n",
       "1  1.573070                   0.0           620.250024        0     26.590354   \n",
       "2  2.810989                   0.0           624.100785        0     27.489738   \n",
       "3  0.872496                   0.0           625.751111        0     27.255089   \n",
       "4  1.818831                   0.0           628.618369        0     26.489366   \n",
       "\n",
       "   appear_time_error  appear_time_time eeg_pulse_time  right_coherence  \\\n",
       "0                0.0        614.707281           None             0.06   \n",
       "1                0.0        618.666397           None             0.24   \n",
       "2                0.0        621.275211           None             0.00   \n",
       "3                0.0        624.850929           None             0.00   \n",
       "4                0.0        626.784654           None             0.00   \n",
       "\n",
       "   left_coherence  press_time_error  press_time_time incorrect_resp pressed  \\\n",
       "0            0.00          0.000510       617.648896              1       1   \n",
       "1            0.30          0.000475       620.239467              4       4   \n",
       "2            0.00          0.000544       624.086200              1       4   \n",
       "3            0.24          0.001483       625.723425              4       1   \n",
       "4            0.12          0.000755       628.603485              4       1   \n",
       "\n",
       "     log_time correct_resp  correct fmri_tr_time  log_num  \n",
       "0  618.148896            4    False         None        0  \n",
       "1  620.739467            1    False         None        0  \n",
       "2  624.586200            4     True         None        0  \n",
       "3  626.223425            1     True         None        0  \n",
       "4  629.103485            1     True         None        0  "
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dl = log2dl('log_MD_0.slog')\n",
    "df = pd.DataFrame(dl)\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# A quick summary\n",
    "\n",
    "- You can use the `describe` method to get a quick summary of your data frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rt</th>\n",
       "      <th>disappear_time_error</th>\n",
       "      <th>disappear_time_time</th>\n",
       "      <th>run_num</th>\n",
       "      <th>refresh_rate</th>\n",
       "      <th>appear_time_error</th>\n",
       "      <th>appear_time_time</th>\n",
       "      <th>right_coherence</th>\n",
       "      <th>left_coherence</th>\n",
       "      <th>press_time_error</th>\n",
       "      <th>press_time_time</th>\n",
       "      <th>log_time</th>\n",
       "      <th>log_num</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.0</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.0</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.000000</td>\n",
       "      <td>248.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.804588</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1481.126965</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>26.667593</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1480.307270</td>\n",
       "      <td>0.150000</td>\n",
       "      <td>0.150000</td>\n",
       "      <td>0.001204</td>\n",
       "      <td>1481.111858</td>\n",
       "      <td>1481.611858</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.383854</td>\n",
       "      <td>0.0</td>\n",
       "      <td>601.187932</td>\n",
       "      <td>1.120295</td>\n",
       "      <td>0.854239</td>\n",
       "      <td>0.0</td>\n",
       "      <td>601.233562</td>\n",
       "      <td>0.104552</td>\n",
       "      <td>0.104552</td>\n",
       "      <td>0.005860</td>\n",
       "      <td>601.187790</td>\n",
       "      <td>601.187790</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.252737</td>\n",
       "      <td>0.0</td>\n",
       "      <td>617.657840</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>21.059205</td>\n",
       "      <td>0.0</td>\n",
       "      <td>614.707281</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000430</td>\n",
       "      <td>617.648896</td>\n",
       "      <td>618.148896</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.558839</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1008.340282</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>26.397669</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1007.321325</td>\n",
       "      <td>0.060000</td>\n",
       "      <td>0.060000</td>\n",
       "      <td>0.000497</td>\n",
       "      <td>1008.328116</td>\n",
       "      <td>1008.828116</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.678502</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1532.049636</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>26.717780</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1530.628267</td>\n",
       "      <td>0.150000</td>\n",
       "      <td>0.150000</td>\n",
       "      <td>0.000519</td>\n",
       "      <td>1532.033580</td>\n",
       "      <td>1532.533580</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.926857</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2005.156651</td>\n",
       "      <td>2.250000</td>\n",
       "      <td>27.143314</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2004.558631</td>\n",
       "      <td>0.240000</td>\n",
       "      <td>0.240000</td>\n",
       "      <td>0.000547</td>\n",
       "      <td>2005.143065</td>\n",
       "      <td>2005.643065</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2.941615</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2243.141804</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>28.609071</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2241.983237</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>0.069283</td>\n",
       "      <td>2243.130664</td>\n",
       "      <td>2243.630664</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               rt  disappear_time_error  disappear_time_time     run_num  \\\n",
       "count  248.000000                 248.0           248.000000  248.000000   \n",
       "mean     0.804588                   0.0          1481.126965    1.500000   \n",
       "std      0.383854                   0.0           601.187932    1.120295   \n",
       "min      0.252737                   0.0           617.657840    0.000000   \n",
       "25%      0.558839                   0.0          1008.340282    0.750000   \n",
       "50%      0.678502                   0.0          1532.049636    1.500000   \n",
       "75%      0.926857                   0.0          2005.156651    2.250000   \n",
       "max      2.941615                   0.0          2243.141804    3.000000   \n",
       "\n",
       "       refresh_rate  appear_time_error  appear_time_time  right_coherence  \\\n",
       "count    248.000000              248.0        248.000000       248.000000   \n",
       "mean      26.667593                0.0       1480.307270         0.150000   \n",
       "std        0.854239                0.0        601.233562         0.104552   \n",
       "min       21.059205                0.0        614.707281         0.000000   \n",
       "25%       26.397669                0.0       1007.321325         0.060000   \n",
       "50%       26.717780                0.0       1530.628267         0.150000   \n",
       "75%       27.143314                0.0       2004.558631         0.240000   \n",
       "max       28.609071                0.0       2241.983237         0.300000   \n",
       "\n",
       "       left_coherence  press_time_error  press_time_time     log_time  log_num  \n",
       "count      248.000000        248.000000       248.000000   248.000000    248.0  \n",
       "mean         0.150000          0.001204      1481.111858  1481.611858      0.0  \n",
       "std          0.104552          0.005860       601.187790   601.187790      0.0  \n",
       "min          0.000000          0.000430       617.648896   618.148896      0.0  \n",
       "25%          0.060000          0.000497      1008.328116  1008.828116      0.0  \n",
       "50%          0.150000          0.000519      1532.033580  1532.533580      0.0  \n",
       "75%          0.240000          0.000547      2005.143065  2005.643065      0.0  \n",
       "max          0.300000          0.069283      2243.130664  2243.630664      0.0  "
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Some data clean-up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['rt', 'disappear_time_error', 'disappear_time_time', 'run_num',\n",
       "       'refresh_rate', 'appear_time_error', 'appear_time_time',\n",
       "       'eeg_pulse_time', 'right_coherence', 'left_coherence',\n",
       "       'press_time_error', 'press_time_time', 'incorrect_resp', 'pressed',\n",
       "       'log_time', 'correct_resp', 'correct', 'fmri_tr_time', 'log_num'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>run_num</th>\n",
       "      <th>trial_num</th>\n",
       "      <th>rt</th>\n",
       "      <th>log_rt</th>\n",
       "      <th>coh_diff</th>\n",
       "      <th>right_coherence</th>\n",
       "      <th>left_coherence</th>\n",
       "      <th>appear_time_time</th>\n",
       "      <th>press_time_error</th>\n",
       "      <th>press_time_time</th>\n",
       "      <th>incorrect_resp</th>\n",
       "      <th>pressed</th>\n",
       "      <th>correct_resp</th>\n",
       "      <th>correct</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2.941615</td>\n",
       "      <td>1.078959</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.00</td>\n",
       "      <td>614.707281</td>\n",
       "      <td>0.000510</td>\n",
       "      <td>617.648896</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.573070</td>\n",
       "      <td>0.453029</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.30</td>\n",
       "      <td>618.666397</td>\n",
       "      <td>0.000475</td>\n",
       "      <td>620.239467</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.810989</td>\n",
       "      <td>1.033537</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>621.275211</td>\n",
       "      <td>0.000544</td>\n",
       "      <td>624.086200</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.872496</td>\n",
       "      <td>-0.136397</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.24</td>\n",
       "      <td>624.850929</td>\n",
       "      <td>0.001483</td>\n",
       "      <td>625.723425</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.818831</td>\n",
       "      <td>0.598194</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>626.784654</td>\n",
       "      <td>0.000755</td>\n",
       "      <td>628.603485</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   run_num  trial_num        rt    log_rt  coh_diff  right_coherence  \\\n",
       "0        0          0  2.941615  1.078959      0.06             0.06   \n",
       "1        0          1  1.573070  0.453029      0.06             0.24   \n",
       "2        0          2  2.810989  1.033537      0.00             0.00   \n",
       "3        0          3  0.872496 -0.136397      0.24             0.00   \n",
       "4        0          4  1.818831  0.598194      0.12             0.00   \n",
       "\n",
       "   left_coherence  appear_time_time  press_time_error  press_time_time  \\\n",
       "0            0.00        614.707281          0.000510       617.648896   \n",
       "1            0.30        618.666397          0.000475       620.239467   \n",
       "2            0.00        621.275211          0.000544       624.086200   \n",
       "3            0.24        624.850929          0.001483       625.723425   \n",
       "4            0.12        626.784654          0.000755       628.603485   \n",
       "\n",
       "  incorrect_resp pressed correct_resp  correct  \n",
       "0              1       1            4    False  \n",
       "1              4       4            1    False  \n",
       "2              1       4            4     True  \n",
       "3              4       1            1     True  \n",
       "4              4       1            1     True  "
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let's add a new column for the absolute value of the coherence difference\n",
    "df['coh_diff'] = np.abs(df['right_coherence'] - df['left_coherence'])\n",
    "\n",
    "# and make a log rt\n",
    "df['log_rt'] = np.log(df['rt'])\n",
    "\n",
    "df['trial_num'] = np.arange(len(df))\n",
    "\n",
    "df = df[['run_num', 'trial_num', 'rt', 'log_rt', 'coh_diff',   \n",
    "         'right_coherence', 'left_coherence', 'appear_time_time',\n",
    "         'press_time_error', 'press_time_time', 'incorrect_resp', 'pressed',\n",
    "         'correct_resp', 'correct']]\n",
    "\n",
    "# show it\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Grouping data\n",
    "\n",
    "- The `groupby` method allows you to create different groupings of your data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['run_num', 'trial_num', 'rt', 'log_rt', 'coh_diff', 'right_coherence',\n",
       "       'left_coherence', 'appear_time_time', 'press_time_error',\n",
       "       'press_time_time', 'incorrect_resp', 'pressed', 'correct_resp',\n",
       "       'correct'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "coh_diff\n",
       "0.00    48\n",
       "0.06    40\n",
       "0.12    64\n",
       "0.18    48\n",
       "0.24    32\n",
       "0.30    16\n",
       "Name: correct, dtype: int64"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('coh_diff')['correct'].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0 0.6041666666666666\n",
      "0.06 0.75\n",
      "0.12 0.78125\n",
      "0.18 0.9583333333333334\n",
      "0.24 0.90625\n",
      "0.3 1.0\n"
     ]
    }
   ],
   "source": [
    "ucd = df['coh_diff'].unique()\n",
    "ucd.sort()\n",
    "for cd in ucd:\n",
    "    print(cd, df.loc[df['coh_diff']==cd, 'correct'].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='coh_diff'>"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# let's look at performance as a function of coh_diff\n",
    "df.groupby('coh_diff')['correct'].mean().plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='coh_diff'>"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# how about rts\n",
    "df.groupby('coh_diff')['rt'].mean().plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='coh_diff'>"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# you can group by more than one column\n",
    "df.groupby(['coh_diff', 'correct'])['rt'].mean().unstack().plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "coh_diff  correct\n",
       "0.00      False      0.829746\n",
       "          True       0.949310\n",
       "0.06      False      1.036445\n",
       "          True       0.861958\n",
       "0.12      False      0.783893\n",
       "          True       0.849955\n",
       "0.18      False      0.884447\n",
       "          True       0.682256\n",
       "0.24      False      0.834226\n",
       "          True       0.701371\n",
       "0.30      True       0.659511\n",
       "Name: rt, dtype: float64"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# you can group by more than one column\n",
    "df.groupby(['coh_diff', 'correct'])['rt'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "correct_resp\n",
       "1   -0.323040\n",
       "4   -0.277191\n",
       "Name: log_rt, dtype: float64"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# is there a speed bias?\n",
    "df.groupby(['correct_resp'])['log_rt'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "correct_resp\n",
       "1    0.879032\n",
       "4    0.733871\n",
       "Name: correct, dtype: float64"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# is there an accuracy bias?\n",
    "df.groupby(['correct_resp'])['correct'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Assignment before next class\n",
    "\n",
    "- Your SMILE experiment is due by Thursday next week.\n",
    "\n",
    "### See you next week!!!"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "smiledev",
   "language": "python",
   "name": "smiledev"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  },
  "rise": {
   "scroll": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}