bars_and_whiskers.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A remake of the bar-and-whisker plot of statistical significance for the RDA paper based on RDA_bars_with_whiskers, but rewritten in much better style and to take advantage of more efficient loading of data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from matplotlib.font_manager import FontProperties\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from shutil import copy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data_file = 'Data/decadal_stats_tau.hdf'\n",
    "results_df = pd.read_hdf(data_file)\n",
    "results_df = results_df.sort_index().sort_index(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Period</th>\n",
       "      <th colspan=\"2\" halign=\"left\">all</th>\n",
       "      <th colspan=\"2\" halign=\"left\">fall</th>\n",
       "      <th colspan=\"2\" halign=\"left\">meiyu</th>\n",
       "      <th colspan=\"2\" halign=\"left\">post-meiyu</th>\n",
       "      <th colspan=\"2\" halign=\"left\">pre-meiyu</th>\n",
       "      <th colspan=\"2\" halign=\"left\">spring</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Years</th>\n",
       "      <th>8007_5179</th>\n",
       "      <th>9407_8093</th>\n",
       "      <th>8007_5179</th>\n",
       "      <th>9407_8093</th>\n",
       "      <th>8007_5179</th>\n",
       "      <th>9407_8093</th>\n",
       "      <th>8007_5179</th>\n",
       "      <th>9407_8093</th>\n",
       "      <th>8007_5179</th>\n",
       "      <th>9407_8093</th>\n",
       "      <th>8007_5179</th>\n",
       "      <th>9407_8093</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Metric</th>\n",
       "      <th>Stat</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"6\" valign=\"top\">frequency</th>\n",
       "      <th>diff_p2p1</th>\n",
       "      <td>-0.00489237</td>\n",
       "      <td>0.000782779</td>\n",
       "      <td>0.0138612</td>\n",
       "      <td>-0.00759878</td>\n",
       "      <td>-0.0243227</td>\n",
       "      <td>-0.00357143</td>\n",
       "      <td>-0.000607328</td>\n",
       "      <td>0.0136986</td>\n",
       "      <td>-0.0601293</td>\n",
       "      <td>0.00178571</td>\n",
       "      <td>0.0119317</td>\n",
       "      <td>-0.0433255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_p1</th>\n",
       "      <td>0.386301</td>\n",
       "      <td>0.381018</td>\n",
       "      <td>0.260455</td>\n",
       "      <td>0.278116</td>\n",
       "      <td>0.668966</td>\n",
       "      <td>0.646429</td>\n",
       "      <td>0.421351</td>\n",
       "      <td>0.413894</td>\n",
       "      <td>0.591379</td>\n",
       "      <td>0.530357</td>\n",
       "      <td>0.465235</td>\n",
       "      <td>0.498829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_p2</th>\n",
       "      <td>0.381409</td>\n",
       "      <td>0.3818</td>\n",
       "      <td>0.274316</td>\n",
       "      <td>0.270517</td>\n",
       "      <td>0.644643</td>\n",
       "      <td>0.642857</td>\n",
       "      <td>0.420744</td>\n",
       "      <td>0.427593</td>\n",
       "      <td>0.53125</td>\n",
       "      <td>0.532143</td>\n",
       "      <td>0.477166</td>\n",
       "      <td>0.455504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pval</th>\n",
       "      <td>0.303991</td>\n",
       "      <td>0.522966</td>\n",
       "      <td>0.71668</td>\n",
       "      <td>0.413541</td>\n",
       "      <td>0.193552</td>\n",
       "      <td>0.464823</td>\n",
       "      <td>0.488811</td>\n",
       "      <td>0.671326</td>\n",
       "      <td>0.02024</td>\n",
       "      <td>0.516886</td>\n",
       "      <td>0.690855</td>\n",
       "      <td>0.102287</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std_p1</th>\n",
       "      <td>0.00669284</td>\n",
       "      <td>0.00960763</td>\n",
       "      <td>0.0168119</td>\n",
       "      <td>0.0247029</td>\n",
       "      <td>0.01954</td>\n",
       "      <td>0.0285706</td>\n",
       "      <td>0.0151769</td>\n",
       "      <td>0.0217882</td>\n",
       "      <td>0.0204117</td>\n",
       "      <td>0.0298256</td>\n",
       "      <td>0.0167714</td>\n",
       "      <td>0.0241966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std_p2</th>\n",
       "      <td>0.00679496</td>\n",
       "      <td>0.00961141</td>\n",
       "      <td>0.0173935</td>\n",
       "      <td>0.024491</td>\n",
       "      <td>0.0202254</td>\n",
       "      <td>0.0286351</td>\n",
       "      <td>0.0154425</td>\n",
       "      <td>0.0218855</td>\n",
       "      <td>0.0210875</td>\n",
       "      <td>0.0298189</td>\n",
       "      <td>0.0170918</td>\n",
       "      <td>0.0241007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"6\" valign=\"top\">intensity</th>\n",
       "      <th>diff_p2p1</th>\n",
       "      <td>0.36322</td>\n",
       "      <td>0.674041</td>\n",
       "      <td>0.753374</td>\n",
       "      <td>1.39691</td>\n",
       "      <td>0.315293</td>\n",
       "      <td>2.66856</td>\n",
       "      <td>0.184915</td>\n",
       "      <td>-0.410096</td>\n",
       "      <td>0.362703</td>\n",
       "      <td>1.32048</td>\n",
       "      <td>0.828424</td>\n",
       "      <td>0.0227796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_p1</th>\n",
       "      <td>23.2812</td>\n",
       "      <td>23.307</td>\n",
       "      <td>20.2594</td>\n",
       "      <td>20.3311</td>\n",
       "      <td>28.1122</td>\n",
       "      <td>27.0932</td>\n",
       "      <td>25.459</td>\n",
       "      <td>25.8524</td>\n",
       "      <td>25.392</td>\n",
       "      <td>25.0903</td>\n",
       "      <td>19.7509</td>\n",
       "      <td>20.5685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_p2</th>\n",
       "      <td>23.6445</td>\n",
       "      <td>23.981</td>\n",
       "      <td>21.0128</td>\n",
       "      <td>21.728</td>\n",
       "      <td>28.4275</td>\n",
       "      <td>29.7618</td>\n",
       "      <td>25.6439</td>\n",
       "      <td>25.4423</td>\n",
       "      <td>25.7547</td>\n",
       "      <td>26.4108</td>\n",
       "      <td>20.5794</td>\n",
       "      <td>20.5913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pval</th>\n",
       "      <td>0.945305</td>\n",
       "      <td>0.980302</td>\n",
       "      <td>0.853715</td>\n",
       "      <td>0.908009</td>\n",
       "      <td>0.714829</td>\n",
       "      <td>0.9997</td>\n",
       "      <td>0.647035</td>\n",
       "      <td>0.277372</td>\n",
       "      <td>0.756724</td>\n",
       "      <td>0.952405</td>\n",
       "      <td>0.984302</td>\n",
       "      <td>0.516348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std_p1</th>\n",
       "      <td>0.155693</td>\n",
       "      <td>0.21969</td>\n",
       "      <td>0.463724</td>\n",
       "      <td>0.691498</td>\n",
       "      <td>0.378743</td>\n",
       "      <td>0.50701</td>\n",
       "      <td>0.348549</td>\n",
       "      <td>0.495789</td>\n",
       "      <td>0.354868</td>\n",
       "      <td>0.515294</td>\n",
       "      <td>0.250295</td>\n",
       "      <td>0.384375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std_p2</th>\n",
       "      <td>0.163406</td>\n",
       "      <td>0.241994</td>\n",
       "      <td>0.510015</td>\n",
       "      <td>0.794565</td>\n",
       "      <td>0.388765</td>\n",
       "      <td>0.58523</td>\n",
       "      <td>0.345229</td>\n",
       "      <td>0.482455</td>\n",
       "      <td>0.397652</td>\n",
       "      <td>0.601009</td>\n",
       "      <td>0.287599</td>\n",
       "      <td>0.425687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"6\" valign=\"top\">latitude</th>\n",
       "      <th>diff_p2p1</th>\n",
       "      <td>-0.249351</td>\n",
       "      <td>-0.163582</td>\n",
       "      <td>0.233537</td>\n",
       "      <td>0.74348</td>\n",
       "      <td>-0.178086</td>\n",
       "      <td>-1.09609</td>\n",
       "      <td>-0.605018</td>\n",
       "      <td>-0.390166</td>\n",
       "      <td>-0.14285</td>\n",
       "      <td>-0.271715</td>\n",
       "      <td>-0.277941</td>\n",
       "      <td>0.321221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_p1</th>\n",
       "      <td>28.7349</td>\n",
       "      <td>28.5675</td>\n",
       "      <td>29.0985</td>\n",
       "      <td>28.9692</td>\n",
       "      <td>29.6035</td>\n",
       "      <td>29.9735</td>\n",
       "      <td>30.1928</td>\n",
       "      <td>29.7861</td>\n",
       "      <td>27.4985</td>\n",
       "      <td>27.4923</td>\n",
       "      <td>27.6092</td>\n",
       "      <td>27.1786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_p2</th>\n",
       "      <td>28.4856</td>\n",
       "      <td>28.4039</td>\n",
       "      <td>29.332</td>\n",
       "      <td>29.7127</td>\n",
       "      <td>29.4254</td>\n",
       "      <td>28.8774</td>\n",
       "      <td>29.5878</td>\n",
       "      <td>29.396</td>\n",
       "      <td>27.3556</td>\n",
       "      <td>27.2206</td>\n",
       "      <td>27.3312</td>\n",
       "      <td>27.4998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pval</th>\n",
       "      <td>0.00309969</td>\n",
       "      <td>0.106289</td>\n",
       "      <td>0.784422</td>\n",
       "      <td>0.972503</td>\n",
       "      <td>0.205879</td>\n",
       "      <td>0.00029997</td>\n",
       "      <td>0.00489951</td>\n",
       "      <td>0.111589</td>\n",
       "      <td>0.241276</td>\n",
       "      <td>0.19668</td>\n",
       "      <td>0.0227977</td>\n",
       "      <td>0.945005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std_p1</th>\n",
       "      <td>0.0650961</td>\n",
       "      <td>0.09196</td>\n",
       "      <td>0.208553</td>\n",
       "      <td>0.25641</td>\n",
       "      <td>0.157057</td>\n",
       "      <td>0.21946</td>\n",
       "      <td>0.168509</td>\n",
       "      <td>0.230248</td>\n",
       "      <td>0.132586</td>\n",
       "      <td>0.227141</td>\n",
       "      <td>0.100867</td>\n",
       "      <td>0.137214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std_p2</th>\n",
       "      <td>0.0637926</td>\n",
       "      <td>0.0904314</td>\n",
       "      <td>0.196724</td>\n",
       "      <td>0.293003</td>\n",
       "      <td>0.154459</td>\n",
       "      <td>0.214518</td>\n",
       "      <td>0.164343</td>\n",
       "      <td>0.228764</td>\n",
       "      <td>0.154209</td>\n",
       "      <td>0.210026</td>\n",
       "      <td>0.100068</td>\n",
       "      <td>0.143198</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Period                      all                    fall              \\\n",
       "Years                 8007_5179    9407_8093  8007_5179   9407_8093   \n",
       "Metric    Stat                                                        \n",
       "frequency diff_p2p1 -0.00489237  0.000782779  0.0138612 -0.00759878   \n",
       "          mean_p1      0.386301     0.381018   0.260455    0.278116   \n",
       "          mean_p2      0.381409       0.3818   0.274316    0.270517   \n",
       "          pval         0.303991     0.522966    0.71668    0.413541   \n",
       "          std_p1     0.00669284   0.00960763  0.0168119   0.0247029   \n",
       "          std_p2     0.00679496   0.00961141  0.0173935    0.024491   \n",
       "intensity diff_p2p1     0.36322     0.674041   0.753374     1.39691   \n",
       "          mean_p1       23.2812       23.307    20.2594     20.3311   \n",
       "          mean_p2       23.6445       23.981    21.0128      21.728   \n",
       "          pval         0.945305     0.980302   0.853715    0.908009   \n",
       "          std_p1       0.155693      0.21969   0.463724    0.691498   \n",
       "          std_p2       0.163406     0.241994   0.510015    0.794565   \n",
       "latitude  diff_p2p1   -0.249351    -0.163582   0.233537     0.74348   \n",
       "          mean_p1       28.7349      28.5675    29.0985     28.9692   \n",
       "          mean_p2       28.4856      28.4039     29.332     29.7127   \n",
       "          pval       0.00309969     0.106289   0.784422    0.972503   \n",
       "          std_p1      0.0650961      0.09196   0.208553     0.25641   \n",
       "          std_p2      0.0637926    0.0904314   0.196724    0.293003   \n",
       "\n",
       "Period                   meiyu               post-meiyu             pre-meiyu  \\\n",
       "Years                8007_5179   9407_8093    8007_5179  9407_8093  8007_5179   \n",
       "Metric    Stat                                                                  \n",
       "frequency diff_p2p1 -0.0243227 -0.00357143 -0.000607328  0.0136986 -0.0601293   \n",
       "          mean_p1     0.668966    0.646429     0.421351   0.413894   0.591379   \n",
       "          mean_p2     0.644643    0.642857     0.420744   0.427593    0.53125   \n",
       "          pval        0.193552    0.464823     0.488811   0.671326    0.02024   \n",
       "          std_p1       0.01954   0.0285706    0.0151769  0.0217882  0.0204117   \n",
       "          std_p2     0.0202254   0.0286351    0.0154425  0.0218855  0.0210875   \n",
       "intensity diff_p2p1   0.315293     2.66856     0.184915  -0.410096   0.362703   \n",
       "          mean_p1      28.1122     27.0932       25.459    25.8524     25.392   \n",
       "          mean_p2      28.4275     29.7618      25.6439    25.4423    25.7547   \n",
       "          pval        0.714829      0.9997     0.647035   0.277372   0.756724   \n",
       "          std_p1      0.378743     0.50701     0.348549   0.495789   0.354868   \n",
       "          std_p2      0.388765     0.58523     0.345229   0.482455   0.397652   \n",
       "latitude  diff_p2p1  -0.178086    -1.09609    -0.605018  -0.390166   -0.14285   \n",
       "          mean_p1      29.6035     29.9735      30.1928    29.7861    27.4985   \n",
       "          mean_p2      29.4254     28.8774      29.5878     29.396    27.3556   \n",
       "          pval        0.205879  0.00029997   0.00489951   0.111589   0.241276   \n",
       "          std_p1      0.157057     0.21946     0.168509   0.230248   0.132586   \n",
       "          std_p2      0.154459    0.214518     0.164343   0.228764   0.154209   \n",
       "\n",
       "Period                              spring             \n",
       "Years                 9407_8093  8007_5179  9407_8093  \n",
       "Metric    Stat                                         \n",
       "frequency diff_p2p1  0.00178571  0.0119317 -0.0433255  \n",
       "          mean_p1      0.530357   0.465235   0.498829  \n",
       "          mean_p2      0.532143   0.477166   0.455504  \n",
       "          pval         0.516886   0.690855   0.102287  \n",
       "          std_p1      0.0298256  0.0167714  0.0241966  \n",
       "          std_p2      0.0298189  0.0170918  0.0241007  \n",
       "intensity diff_p2p1     1.32048   0.828424  0.0227796  \n",
       "          mean_p1       25.0903    19.7509    20.5685  \n",
       "          mean_p2       26.4108    20.5794    20.5913  \n",
       "          pval         0.952405   0.984302   0.516348  \n",
       "          std_p1       0.515294   0.250295   0.384375  \n",
       "          std_p2       0.601009   0.287599   0.425687  \n",
       "latitude  diff_p2p1   -0.271715  -0.277941   0.321221  \n",
       "          mean_p1       27.4923    27.6092    27.1786  \n",
       "          mean_p2       27.2206    27.3312    27.4998  \n",
       "          pval          0.19668  0.0227977   0.945005  \n",
       "          std_p1       0.227141   0.100867   0.137214  \n",
       "          std_p2       0.210026   0.100068   0.143198  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#customized script that adds overbars to indicate time periods, and then on top of that a p-value for the significance of the change.\n",
    "def add_pvals(fig,ax,rects_in,errs_list,pvals_list):\n",
    "    \n",
    "    #reorders rectangles so that we're dealing with group of 3 rectangles for each time period at same time\n",
    "    zipped_rects = zip(rects_in[0],rects_in[1],rects_in[2])\n",
    "    \n",
    "    #get y-axis limits\n",
    "    yax = ax.get_ylim()\n",
    "    yhgt = yax[1]-yax[0]\n",
    "    \n",
    "    #add p-values sequentially to each group of 3 columns\n",
    "    for rects,errs,pvals in zip(zipped_rects,errs_list,pvals_list):\n",
    "        \n",
    "        ##each rects group contains 3 bars - 1st bar is for '51-'79 and is wider, 2nd is '80-'93 and 3rd is '94-'07\n",
    "        heights = [ rect.get_height() + err for rect,err in zip(rects,errs) ]\n",
    "        max_height = max(heights[0],heights[1],heights[2])\n",
    "\n",
    "        ## figure out x and y coordinates of brackets that are placed above bars\n",
    "        xx_l = [rects[0].get_x() + rects[0].get_width()/2, rects[1].get_x() + rects[1].get_width()/2]\n",
    "        xx_r = [ xx_l[0] + rects[0].get_width(), xx_l[1] + rects[1].get_width() ]\n",
    "        yy = [max_height + .01*yhgt, max_height + .13*yhgt]\n",
    "        bracket_widths = [1.25,.625]\n",
    "        \n",
    "        for x_l, x_r, y, bwidth, pval in zip(xx_l, xx_r, yy, bracket_widths, pvals):\n",
    "            ##add brackets\n",
    "            ax.annotate(\"\", xy=(x_l, y + .015*yhgt), xytext=(x_l, y + .025*yhgt), \n",
    "                fontsize=15, ha='center', va='bottom',\n",
    "                arrowprops=dict(arrowstyle='-[, widthB='+str(bwidth)+', lengthB=.5', lw=1.0))\n",
    "\n",
    "            ax.annotate(\"\", xy=(x_r, y + .015*yhgt), xytext=(x_r, y + .025*yhgt), \n",
    "                fontsize=15, ha='center', va='bottom',\n",
    "                arrowprops=dict(arrowstyle='-[, widthB='+str(bwidth)+', lengthB=.5', lw=1.0))\n",
    "\n",
    "            #add extra line\n",
    "            ax.plot([x_l, (x_l+x_r)/2], [y + .022*yhgt, y + .045*yhgt], 'k-', lw=1)\n",
    "            ax.plot([x_r, (x_l+x_r)/2], [y + .022*yhgt, y + .045*yhgt], 'k-', lw=1)\n",
    "\n",
    "            #add p-value\n",
    "            font = FontProperties()\n",
    "            font.set_size('large')\n",
    "            \n",
    "            if abs(pval-.5) > .475:\n",
    "                font.set_weight('bold')\n",
    "                if abs(pval-.5) > .495:\n",
    "                    font.set_style('italic')\n",
    "            if pval > .999:\n",
    "                message = 'p>.999'\n",
    "            elif pval < .001:\n",
    "                message = 'p<.001'\n",
    "            else:\n",
    "                message = \"p={0:.3f}\".format(round(pval,3))\n",
    "\n",
    "            ax.text((x_l+x_r)/2, y + .06*yhgt, message, horizontalalignment = 'center', fontproperties = font)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "N = 6               # number of data entries\n",
    "ind = np.arange(N)              # the x locations for the groups\n",
    "width1 = 0.35                   # bar width\n",
    "width2 = 0.175"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.text.Text at 0x110a663c8>,\n",
       " <matplotlib.text.Text at 0x110a756d8>,\n",
       " <matplotlib.text.Text at 0x1141f7b70>,\n",
       " <matplotlib.text.Text at 0x1141fe588>,\n",
       " <matplotlib.text.Text at 0x1141fef60>,\n",
       " <matplotlib.text.Text at 0x114208978>,\n",
       " <matplotlib.text.Text at 0x11420c0b8>,\n",
       " <matplotlib.text.Text at 0x11420ca90>,\n",
       " <matplotlib.text.Text at 0x1142134a8>,\n",
       " <matplotlib.text.Text at 0x114213e80>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(3,1,figsize=(12,18), facecolor='w')\n",
    "rects = []\n",
    "\n",
    "## Parameters for subplots\n",
    "colors = ['MediumSlateBlue','Tomato','Gold']\n",
    "metrics = ['frequency','latitude','intensity']\n",
    "my_error_kw={'ecolor':'Black', 'linewidth':1, 'capsize':2}\n",
    "periods = ['p1','p1','p2']\n",
    "seasons = ['spring','pre-meiyu','meiyu','post-meiyu','fall','all']\n",
    "subplot_titles = ['a) Frontal Rain Event (FRE) Frequency (%)','b) Mean FRE Latitude ($^{\\circ}$N)',\\\n",
    "                  'c) Mean FRE Daily Accumulation (mm day$^{-1}$)']\n",
    "widths = [width1, width2, width2]\n",
    "xpositions = [ind-(width1/2),ind+(width2/2),ind+(3*width2/2)]\n",
    "xtick_labels = ('Spring', 'Pre-Meiyu', 'Meiyu', 'Post-Meiyu', 'Fall','Full Year')\n",
    "yaxis_labels = ['Frequency (%)','Latitude ($^{\\circ}$N)','Daily Accumulation (mm day$^{-1}$)']\n",
    "year_tags = ['8007_5179','9407_8093','9407_8093']\n",
    "yranges = [[0,.9],[26.8,31.6],[19,34]]\n",
    "\n",
    "\n",
    "for ax,met,title,ylabel,yrange in zip(axes,metrics,subplot_titles,yaxis_labels,yranges):\n",
    "    \n",
    "    #list of all 3 bar plots for a given subpanel\n",
    "    sub_rects = []\n",
    "    \n",
    "    ## Plot set of bars for each time period\n",
    "    ## the [seasons] toggle organizes values in correct order\n",
    "    for (color, p, width, xpos, yr_tag) in zip(colors, periods, widths, xpositions, year_tags):\n",
    "        vals = results_df.loc[(met,'mean_'+p),(slice(None),yr_tag)][seasons].values\n",
    "        errs = results_df.loc[(met,'std_'+p),(slice(None),yr_tag)][seasons].values\n",
    "        sub_rects.append(ax.bar(xpos,vals,width, color=color, yerr=errs, error_kw=my_error_kw,\\\n",
    "                               edgecolor='black'))\n",
    "    \n",
    "    ## Tweak plot formatting\n",
    "    ax.legend((sub_rects[0], sub_rects[1], sub_rects[2]), (\"'51-'79\", \"'80-'93\", \"'94-'07\"),\\\n",
    "              prop={'size': 14})\n",
    "    ax.set_title(title, fontsize=22)\n",
    "    ax.set_xlim([-.5,5.5])\n",
    "    ax.set_xticks(xpositions[0]+width1/2)\n",
    "    ax.set_xticklabels(xtick_labels,\\\n",
    "                        fontsize=16)\n",
    "    #ax.set_ylabel(ylabel, fontsize=16)\n",
    "    ax.set_ylim([yrange[0],yrange[1]])\n",
    "\n",
    "    ## Turn off x-axis ticks\n",
    "    ax.tick_params(\n",
    "    axis='x',          # changes apply to the x-axis\n",
    "    which='both',      # both major and minor ticks are affected\n",
    "    bottom='off',      # ticks along the bottom edge are off\n",
    "    top='off')         # ticks along the top edge are off\n",
    "    \n",
    "    ## Change default size of y-axis tick labels\n",
    "    ax.tick_params(\n",
    "    axis='y',          # changes apply to the x-axis\n",
    "    labelsize=14)         # ticks along the top edge are off\n",
    "    \n",
    "    rects.append(sub_rects)\n",
    "    \n",
    "    ## collect the right collection of errors and p-values necessary for plotting.\n",
    "    sub_errs = []\n",
    "    sub_pvals = []\n",
    "    \n",
    "    for season in seasons:\n",
    "        \n",
    "        ## arrange errors\n",
    "        bar_errs = []\n",
    "        for (p, yr_tag) in zip(periods,year_tags):\n",
    "            bar_errs.append(results_df.loc[(met,'std_'+p),(season,yr_tag)])\n",
    "        \n",
    "        sub_errs.append(bar_errs)\n",
    "        sub_pvals.append(list(results_df.loc[(met,'pval'),(season,slice(None))].values))\n",
    "\n",
    "    #add p-values to top of plots\n",
    "    add_pvals(fig,ax,sub_rects,sub_errs,sub_pvals)\n",
    "  \n",
    "## Change y-axis labels from frequency to percentage in subplot a\n",
    "axes[0].set_yticklabels(np.linspace(0,90,10).astype(int))\n",
    "\n",
    "#plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overwriting existing version of file in RDA manuscript directory...\n",
      "Overwriting existing version of file in RDA Python directory...\n"
     ]
    }
   ],
   "source": [
    "## SAVE FIGURE TO MANUSCRIPT AND ANALYSIS DIRECTORIES\n",
    "figure_name = \"bars.pdf\"\n",
    "\n",
    "savepath = \"/Users/Siwen/Desktop/RDA/Figures/\" + figure_name\n",
    "\n",
    "if os.path.exists(savepath):\n",
    "    print(\"Overwriting existing version of file in RDA manuscript directory...\")\n",
    "    os.remove(savepath)\n",
    "    \n",
    "plt.savefig(savepath, bbox_inches='tight')\n",
    "    \n",
    "copypath = \"/Users/Siwen/RDA/Figures/\" + figure_name\n",
    "    \n",
    "if os.path.exists(copypath):\n",
    "    print(\"Overwriting existing version of file in RDA Python directory...\")\n",
    "    os.remove(copypath)\n",
    "    \n",
    "copy(savepath,copypath)\n",
    "\n",
    "plt.savefig(savepath, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}