tea_demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pathlib\n",
    "\n",
    "import requests\n",
    "\n",
    "import tea"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data\n",
    "\n",
    "data_path = pathlib.Path(\"./datasets/UScrime.csv\")\n",
    "data_path.parent.mkdir(parents = True, exist_ok = True)\n",
    "\n",
    "data_url = \"https://raw.githubusercontent.com/vincentarelbundock/Rdatasets/master/csv/MASS/UScrime.csv\"\n",
    "data_path.write_text(requests.get(data_url).text)\n",
    "\n",
    "tea.data(data_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Declare and annotate the variables of interest\n",
    "variables = [\n",
    "    {\n",
    "        'name' : 'So',\n",
    "        'data type' : 'nominal',\n",
    "        'categories' : ['0', '1']\n",
    "    },\n",
    "    {\n",
    "        'name' : 'Prob',\n",
    "        'data type' : 'ratio',\n",
    "        'range' : [0,1]\n",
    "    }\n",
    "]\n",
    "\n",
    "tea.define_variables(variables)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Running under STRICT mode.\n",
      "\n",
      "This means that user assertions will be checked. Should they fail, Tea will override user assertions.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "assumptions = {\n",
    "    'groups normally distributed': [['So', 'Prob']],\n",
    "    'Type I (False Positive) Error Rate': 0.05,\n",
    "}\n",
    "\n",
    "tea.assume(assumptions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "experimental_design = {\n",
    "                        'study type': 'observational study',\n",
    "                        'contributor variables': 'So',\n",
    "                        'outcome variables': 'Prob',\n",
    "                    }\n",
    "tea.define_study_design(experimental_design)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running under STRICT mode.\n",
      "User asserted property: is_groups_normal is supported by statistical checking. Tea agrees with the user.\n",
      "\n",
      "Currently considering pearson_corr\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property FAILS\n",
      "Testing assumption: is_continuous.\n",
      "Testing assumption: is_normal.\n",
      "Testing assumption: is_normal.\n",
      "\n",
      "Currently considering kendalltau_corr\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous_or_ordinal.\n",
      "Property FAILS\n",
      "Testing assumption: is_continuous_or_ordinal.\n",
      "\n",
      "Currently considering spearman_corr\n",
      "Test is unsat.\n",
      "\n",
      "\n",
      "Currently considering pointbiserial_corr_a\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property holds.\n",
      "Testing assumption: is_normal.\n",
      "Property FAILS\n",
      "Testing assumption: is_categorical.\n",
      "Testing assumption: has_two_categories.\n",
      "Testing assumption: has_equal_variance.\n",
      "\n",
      "Currently considering pointbiserial_corr_b\n",
      "Test is unsat.\n",
      "\n",
      "\n",
      "Currently considering students_t\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: has_one_x.\n",
      "Property holds.\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: has_two_categories.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property holds.\n",
      "Testing assumption: has_equal_variance.\n",
      "Property holds.\n",
      "User asserted property: is_groups_normal.\n",
      "\n",
      "Currently considering welchs_t\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: has_one_x.\n",
      "Property holds.\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: has_two_categories.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property holds.\n",
      "User asserted property: is_groups_normal.\n",
      "\n",
      "Currently considering mannwhitney_u\n",
      "Testing assumption: has_one_x.\n",
      "Property holds.\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: has_two_categories.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous_or_ordinal.\n",
      "Property holds.\n",
      "\n",
      "Currently considering paired_students_t\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: has_one_x.\n",
      "Property holds.\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: has_paired_observations.\n",
      "Property FAILS\n",
      "Testing assumption: is_categorical.\n",
      "Testing assumption: has_two_categories.\n",
      "Testing assumption: is_continuous.\n",
      "User asserted property: is_groups_normal.\n",
      "Testing assumption: has_equal_variance.\n",
      "\n",
      "Currently considering wilcoxon_signed_rank\n",
      "Test is unsat.\n",
      "\n",
      "\n",
      "Currently considering chi_square\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: greater_than_5_freq.\n",
      "Property FAILS\n",
      "Testing assumption: is_categorical.\n",
      "Testing assumption: is_categorical.\n",
      "Testing assumption: has_two_or_more_categories.\n",
      "Testing assumption: has_two_or_more_categories.\n",
      "\n",
      "Currently considering fishers_exact\n",
      "Testing assumption: is_bivariate.\n",
      "Property holds.\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property FAILS\n",
      "Testing assumption: has_two_categories.\n",
      "Testing assumption: has_two_categories.\n",
      "\n",
      "Currently considering f_test\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: has_one_x.\n",
      "Property holds.\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: has_two_or_more_categories.\n",
      "Property holds.\n",
      "User asserted property: is_groups_normal.\n",
      "Testing assumption: has_equal_variance.\n",
      "Property holds.\n",
      "\n",
      "Currently considering kruskall_wallis\n",
      "Testing assumption: has_independent_observations.\n",
      "Property holds.\n",
      "Testing assumption: has_one_x.\n",
      "Property holds.\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: has_two_or_more_categories.\n",
      "Property holds.\n",
      "\n",
      "Currently considering rm_one_way_anova\n",
      "Test is unsat.\n",
      "\n",
      "\n",
      "Currently considering friedman\n",
      "Test is unsat.\n",
      "\n",
      "\n",
      "Currently considering factorial_ANOVA\n",
      "Testing assumption: has_one_y.\n",
      "Property holds.\n",
      "Testing assumption: is_continuous.\n",
      "Property holds.\n",
      "Testing assumption: is_categorical.\n",
      "Property holds.\n",
      "Testing assumption: has_two_or_more_categories.\n",
      "Property holds.\n",
      "User asserted property: is_groups_normal.\n",
      "Testing assumption: has_equal_variance.\n",
      "Property holds.\n",
      "\n",
      "\n",
      "Results:\n",
      "--------------\n",
      "Test: students_t\n",
      "***Test assumptions:\n",
      "Exactly two variables involved in analysis: So, Prob\n",
      "Exactly one explanatory variable: So\n",
      "Exactly one explained variable: Prob\n",
      "Independent (not paired) observations: So\n",
      "Variable is categorical: So\n",
      "Variable has two categories: So\n",
      "Continuous (not categorical) data: Prob\n",
      "Equal variance: So, Prob\n",
      "Groups are normally distributed: So, Prob\n",
      "\n",
      "***Test results:\n",
      "name = Student's T Test\n",
      "test_statistic = 4.20213\n",
      "p_value = 0.00012\n",
      "adjusted_p_value = 0.00006\n",
      "alpha = 0.05\n",
      "dof = 45\n",
      "Effect size:\n",
      "Cohen's d = 1.24262\n",
      "A12 = 0.83669\n",
      "Null hypothesis = There is no difference in means between So = 0 and So = 1 on Prob.\n",
      "Interpretation = t(45) = 4.20213, p = 0.00006. Reject the null hypothesis at alpha = 0.05. The mean of Prob for So = 1 (M=0.06371, SD=0.02251) is significantly greater than the mean for So = 0 (M=0.03851, SD=0.01778). The effect size is Cohen's d = 1.24262, A12 = 0.83669. The effect size is the magnitude of the difference, which gives a holistic view of the results [1].\n",
      "[1] Sullivan, G. M., & Feinn, R. (2012). Using effect size—or why the P value is not enough. Journal of graduate medical education, 4(3), 279-282.\n",
      "\n",
      "Test: welchs_t\n",
      "***Test assumptions:\n",
      "Exactly two variables involved in analysis: So, Prob\n",
      "Exactly one explanatory variable: So\n",
      "Exactly one explained variable: Prob\n",
      "Independent (not paired) observations: So\n",
      "Variable is categorical: So\n",
      "Variable has two categories: So\n",
      "Continuous (not categorical) data: Prob\n",
      "Groups are normally distributed: So, Prob\n",
      "\n",
      "***Test results:\n",
      "name = Welch's T Test\n",
      "test_statistic = 3.89537\n",
      "p_value = 0.00065\n",
      "adjusted_p_value = 0.00033\n",
      "alpha = 0.05\n",
      "dof = 15\n",
      "Effect size:\n",
      "A12 = 0.83669\n",
      "Null hypothesis = There is no difference in means between So = 0 and So = 1 on Prob.\n",
      "Interpretation = t(15) = 3.89537, p = 0.00033. Reject the null hypothesis at alpha = 0.05. The mean of Prob for So = 1 is significantly greater than the mean for So = 0. The effect size is A12 = 0.83669. The effect size is the magnitude of the difference, which gives a holistic view of the results [1].\n",
      "[1] Sullivan, G. M., & Feinn, R. (2012). Using effect size—or why the P value is not enough. Journal of graduate medical education, 4(3), 279-282.\n",
      "\n",
      "Test: mannwhitney_u\n",
      "***Test assumptions:\n",
      "Exactly one explanatory variable: So\n",
      "Exactly one explained variable: Prob\n",
      "Independent (not paired) observations: So\n",
      "Variable is categorical: So\n",
      "Variable has two categories: So\n",
      "Continuous OR ORDINAL (not nominal) data: Prob\n",
      "\n",
      "***Test results:\n",
      "name = Mann Whitney U Test\n",
      "test_statistic = 81.00000\n",
      "p_value = 0.00019\n",
      "adjusted_p_value = 0.00009\n",
      "alpha = 0.05\n",
      "dof = 31\n",
      "Effect size:\n",
      "A12 = 0.83669\n",
      "Null hypothesis = There is no difference in medians between So = 0 and So = 1 on Prob.\n",
      "Interpretation = t(31) = 81.00000, p = 0.00009. Reject the null hypothesis at alpha = 0.05. The median of Prob for So = 1 is significantly greater than the median for So = 0. The effect size is A12 = 0.83669. The effect size is the magnitude of the difference, which gives a holistic view of the results [1].\n",
      "[1] Sullivan, G. M., & Feinn, R. (2012). Using effect size—or why the P value is not enough. Journal of graduate medical education, 4(3), 279-282.\n",
      "\n",
      "Test: f_test\n",
      "***Test assumptions:\n",
      "Independent (not paired) observations: So\n",
      "Exactly one explanatory variable: So\n",
      "Exactly one explained variable: Prob\n",
      "Continuous (not categorical) data: Prob\n",
      "Variable is categorical: So\n",
      "Variable has two or more categories: So\n",
      "Groups are normally distributed: So, Prob\n",
      "Equal variance: So, Prob\n",
      "\n",
      "***Test results:\n",
      "name = F Test\n",
      "test_statistic = 17.65790\n",
      "p_value = 0.00012\n",
      "adjusted_p_value = 0.00012\n",
      "alpha = 0.05\n",
      "dof = 1.0\n",
      "table =             df    sum_sq   mean_sq          F    PR(>F)\n",
      "C(So)      1.0  0.006702  0.006702  17.657903  0.000124\n",
      "Residual  45.0  0.017079  0.000380        NaN       NaN\n",
      "Null hypothesis = The variances of the groups (So = 0, 1) are equal.\n",
      "Interpretation = t(1.0) = 17.65790, p = 0.00012. Reject the null hypothesis at alpha = 0.05. The variances of the groups (So = 0, 1) are not equal.\n",
      "\n",
      "Test: kruskall_wallis\n",
      "***Test assumptions:\n",
      "Independent (not paired) observations: So\n",
      "Exactly one explanatory variable: So\n",
      "Exactly one explained variable: Prob\n",
      "Continuous (not categorical) data: Prob\n",
      "Variable is categorical: So\n",
      "Variable has two or more categories: So\n",
      "\n",
      "***Test results:\n",
      "name = Kruskall Wallis\n",
      "test_statistic = 14.05696\n",
      "p_value = 0.00018\n",
      "adjusted_p_value = 0.00018\n",
      "alpha = 0.05\n",
      "dof = 31\n",
      "Null hypothesis = There is no difference in medians between So = 0, 1 on Prob.\n",
      "Interpretation = t(31) = 14.05696, p = 0.00018. Reject the null hypothesis at alpha = 0.05. There is a difference in medians of Prob for at least one of So = 0, 1.\n",
      "\n",
      "Test: factorial_ANOVA\n",
      "***Test assumptions:\n",
      "Exactly one explained variable: Prob\n",
      "Continuous (not categorical) data: Prob\n",
      "Variable is categorical: So\n",
      "Variable has two or more categories: So\n",
      "Groups are normally distributed: So, Prob\n",
      "Equal variance: So, Prob\n",
      "\n",
      "***Test results:\n",
      "name = Factorial ANOVA\n",
      "test_statistic = 17.65790\n",
      "p_value = 0.00012\n",
      "adjusted_p_value = 0.00012\n",
      "alpha = 0.05\n",
      "dof = 1.0\n",
      "table =             df    sum_sq   mean_sq          F    PR(>F)\n",
      "C(So)      1.0  0.006702  0.006702  17.657903  0.000124\n",
      "Residual  45.0  0.017079  0.000380        NaN       NaN\n",
      "Null hypothesis = There is no difference in means between So = 0, 1 on Prob.\n",
      "Interpretation = t(1.0) = 17.65790, p = 0.00012. Reject the null hypothesis at alpha = 0.05. There is a difference in means of Prob for at least one of So = 0, 1.\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tea.runtimeDataStructures.resultData.ResultData at 0x1c2429f890>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tea.hypothesize(['So', 'Prob'], ['So:1 > 0'])"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}