Report.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "<h3>Description</h3>\n",
    "<p>Increasing the yield effeciency is a requirement that is very difficult to pursue manually, as it requires manual effort and considerable time window to see the results. Hence, we try to build a model (based on DeepLearning) that can carry out the process many times faster and without the involvemnet of any real time queries on the processes. All we require is the data collected,pertaining to parameters and objects , prior to carrying out the processes/unit steps. However, the proposed model can only be applied on a single process for its optimisation</p>\n",
    "\n",
    "<hr>\n",
    "\n",
    "<h3>Methodology proposed</h3>\n",
    "<br>\n",
    "The following steps were proposed to carry out the research on the data\n",
    "<ol>\n",
    "    <li> <b>Finding</b> out the process on which the optimisation has to be applied</li>\n",
    "    <li> <b>Fetching the data</b> from sql table into csv format, (as python is very comfortable with csvs')</li>\n",
    "    <li> <b>Analysing</b> the selected process to find:\n",
    "        <ul>\n",
    "            <li>the mean and the standard deviations for each parameter id\n",
    "            <li>the expected distribution of the parameters for the process (assuming the process parameters to follow a gaussian distribution).\n",
    "            <li>Graphing the results as histograms/line_curves to outline the presence of outliers\n",
    "            <li>from the above insights, finding out outlier data pertaining to the measured values of the given data.\n",
    "        </ul>\n",
    "    </li>\n",
    "    <li> Removing the outliers. normalising the data and other <b>preprocessing</b> techniques to clean data. Finally, learning the effect of input parameters on the ouput.\n",
    "    <li> From the weights and relation obtained varying the input parameters through a range of LSL to USL and <b>find its value for optimal output</b> (best yield)</li>\n",
    "    <li><b>output the final values</b> that should be applied for the process.</li>\n",
    "</ol>\n",
    "<p>The following points will cover the above mentioned methodology in a greater detail.</p>\n",
    "\n",
    "<h3><u>Note: environment</u></h3>\n",
    "<p>the code presumes you have python3 installed on your machine.\n",
    "The following python3 libraries are required to execute the cells.<p>\n",
    "<ul>\n",
    "    <li>pandas</li>\n",
    "    <li>numpy</li>\n",
    "    <li>scipy</li>\n",
    "    <li>matplotlib</li>\n",
    "    </ul>\n",
    "you may simply run <code> pip install requirements.txt</code> to ensure the libraries are installed."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>1. Process for carrying out research</h2>\n",
    "<p> The process chosen to carry out this research was \"Product Flow 1.8V/3.3V w\". As this process was quite useful and had the scope to be imporved our further analysis assumes all other data is only related to this process.</p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>2. Fetching data from sql database</h2>\n",
    "<p> The following command on executing fetches the required data for further analysis. The output of the data is stored in <code>data/data.csv</code> </p>\n",
    "\n",
    "```mysql\n",
    "SELECT A.LotID,A.DeviceID,A.ProcessID , C.OperationID, C.ParameterId, C.MeasuredValue as MV, C.WaferNo, E.UpperHoldValue as USL, E.LowerHoldValue as LSL, E.UpperControlLimit as UCL, E.LowerControlLimit as LCL\n",
    "FROM\n",
    "  LLotStart A\n",
    "  inner join LOperationMeasuredValueRaw C on C.LotID = A.LotID\n",
    "  inner join ParameterIdForRnQA1 B on B.ParameterId  = C.ParameterId\n",
    "  inner join LOperationLimits E on E.OpParameterID = B.ParameterId and E.OperationID = C.OperationID\n",
    "  inner join LOperationParameterUnit F on F.OpParameterId = B.ParameterId and F.OperationID = C.OperationID\n",
    "WHERE A.LotID like 'F%' and ProcessID = 'Product Flow 1.8V/3.3V w'\t\n",
    "order by C.ParameterId, LotID desc\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#-----processing the spacing error in the fetched data------\n",
    "#-----extracting data-------\n",
    "file = open('data/data.csv','r')\n",
    "data = file.read()\n",
    "file.close()\n",
    "\n",
    "#-------formatting data------------\n",
    "data = data.split('\\n')\n",
    "for i in range(0,len(data)):\n",
    "    data[i]=data[i].split(',')\n",
    "    for j in range(len(data[i])):\n",
    "        data[i][j]=data[i][j].strip()\n",
    "data[0][0]='lotid'\n",
    "\n",
    "#----writing to the final file for csv--------\n",
    "file=open('data/processed_data.csv','w')\n",
    "s=''\n",
    "for i in range(0,len(data)-1):\n",
    "    s+=data[i][0]\n",
    "    for j in range(1,len(data[i])):\n",
    "        s+=','+data[i][j]\n",
    "    s+='\\n'\n",
    "file.write(s)\n",
    "file.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>3. Analysing</h2>\n",
    "<p>  The above cell would result in <code>data/processed_data.csv</code>. As mentioned above we follow the steps indicated below to further analyse the so created data.</p>\n",
    "<hr>\n",
    "\n",
    "<h3>Finding mean,standard deviations for each parameter id</h3>\n",
    "<p> The following python snippet does the entire work of finding</p>\n",
    "<p> Some essential points to note:\n",
    "    <ul>\n",
    "        <li> Since the provided data does not contain the target value we assume target value to be around the mean for that will remain same throughout for a given parameter under the process.</li>\n",
    "        <li> The mean calculated is only wrt to the given process so it includes all the sum of all the values of measured value for a given parameter and dividing by total occurances, but the scope of domain is only the process included, as the same parameter might have been used for other process too, but will not be considered in these means. The same goes for the calculated standard deviations.</li>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!/usr/bin/env python3\n",
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Wed Jun 17 21:06:59 2020\n",
    "\n",
    "@author: dipanshu\n",
    "\"\"\"\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "data = pd.read_csv('data/processed_data.csv')\n",
    "\n",
    "parameter_grouped_data = data.groupby('parameterid')\n",
    "mv_stdev = parameter_grouped_data.mv.std()\n",
    "mv_mean = parameter_grouped_data.mv.mean()\n",
    "\n",
    "for i in range(len(data)):\n",
    "    data.at[i,'mean']=mv_mean[data.loc[i]['parameterid']]\n",
    "    data.at[i,'standard_dev']=mv_stdev[data.loc[i]['parameterid']]\n",
    "    data.at[i,'target_value']=int(mv_mean[data.loc[i]['parameterid']])\n",
    "    \n",
    "data.to_csv('data/data_w_mean_std.csv', index=False)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Expected distribution of the parameters for the process</h3>\n",
    "<p> Some essential points to note:\n",
    "    <ul>\n",
    "        <li> we have assumed the data to follow a gaussian distribution for each parameter. The gaussian distribution has a mean and a standard deviation as its parameters. (Also called bell shaped curve)</li>\n",
    "        <li>\n",
    "            The distribution has the following distribution, where mu=mean and sigma=standard_deviation\n",
    "        </li>\n",
    "\n",
    "![gaussian.png](gaussian.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Graphing the results as histograms/line_curves to outline the presence of outliers</h3>\n",
    "<p> The following python snippet does the entire work of graphing the various plots</p>\n",
    "<p> Some essential points to note:\n",
    "    <ul>\n",
    "        <li> Some parameters have a very large value of USL inserted. These values are dummy values and hence do not occur in the provided artist for the graph. the graph covers a range on the x axis from (mean-10*stdev) to (mean+10*stdev)</li>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!/usr/bin/env python3\n",
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Sat Jun 20 11:35:16 2020\n",
    "\n",
    "@author: dipanshu\n",
    "\"\"\"\n",
    "\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import scipy.stats as st\n",
    "\n",
    "data = pd.read_csv('data/data_w_mean_std.csv')\n",
    "\n",
    "parameter_grouped_data = dict(list(data.groupby('parameterid')))\n",
    "parameters = list(parameter_grouped_data.keys())\n",
    "for index,parameter in enumerate(parameters):\n",
    "    plt.figure(figsize=[12,6])\n",
    "    \n",
    "    mean=parameter_grouped_data[parameter]['mean'].mean()\n",
    "    standard_dev=parameter_grouped_data[parameter].standard_dev.mean()\n",
    "    target = parameter_grouped_data[parameter].target_value.mean()\n",
    "    if(mean==target):\n",
    "        target+=1\n",
    "    \n",
    "    if(standard_dev==0):\n",
    "        print('data not enough... aborting for',parameter)\n",
    "        continue\n",
    "    \n",
    "    usl = parameter_grouped_data[parameter].usl.mean()\n",
    "    lsl = parameter_grouped_data[parameter].lsl.mean()\n",
    "    \n",
    "    plt.axvline(usl,linestyle='dashed',label='USL ='+str(usl),color='b')\n",
    "    plt.axvline(lsl,linestyle='dashed',label='LSL ='+str(lsl),color='g')\n",
    "    plt.axvline(mean,linestyle='dashed',label='mean ='+str(mean),color='red')\n",
    "    plt.axvline(target,linestyle='dashed',label='target value ='+str(target),color='yellow')\n",
    "    \n",
    "    n,bins,patches = plt.hist(parameter_grouped_data[parameter].mv, align='mid',\\\n",
    "                              color=(0.4, 0.6, 0.6, 1), edgecolor='black',density=True)\n",
    "    for i in range(len(bins)-1):\n",
    "        bar_val = (bins[i]+bins[i+1])/2\n",
    "        if(bar_val<lsl or bar_val>usl):\n",
    "            patches[i].set_fc('r')\n",
    "    \n",
    "    bell_x = [(bins[i]+bins[i+1])/2 for i in range(len(bins)-1)]\n",
    "    bell_y = n\n",
    "    expected_x = np.linspace(mean-standard_dev*5,mean+standard_dev*5,1000)\n",
    "    expected_y = st.norm(mean,standard_dev).pdf(expected_x)\n",
    "    \n",
    "    plt.plot(bell_x,bell_y,linestyle='dashed',color='black',label='lineCurve')\n",
    "    plt.plot(expected_x,expected_y,color='red',label='expected bell curve')\n",
    "    plt.xlim([mean-10*standard_dev,mean+10*standard_dev])\n",
    "    \n",
    "    plt.legend()\n",
    "    plt.savefig('graphs/'+str(parameter)+'.png')\n",
    "    plt.close()\n",
    "    print('progress: ',round((index/len(parameters))*100,2),'graphed',parameter,'...')\n",
    "\n",
    "print('process has completed... thank you for your patience =)')\n"
   ]
  },
  {
   "attachments": {
    "DICD_NS_Iso_Space_3Sigma.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<p><strong>one of the graphs with correct data looks as follows</strong></p>\n",
    "<p>Key points to note\n",
    "    <ul>\n",
    "        <li> where the x axis is the measured values while the y axis is the distribution (fraction of the occurence of that bin)</li>\n",
    "        <li> the bars in green represent that these are acceptable values, while those in red represent non-acceptable values</li>\n",
    "    </ul>\n",
    "<p>\n",
    "\n",
    "![DICD_NS_Iso_Space_3Sigma.png](attachment:DICD_NS_Iso_Space_3Sigma.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Finding out outlier data pertaining to the measured values of the given data.</h3>\n",
    "<p> The following python snippet does the entire work of outputing the csv containing info related to those outliers</p>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!/usr/bin/env python3\n",
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Sun Jun 21 12:32:11 2020\n",
    "\n",
    "@author: dipanshu\n",
    "\"\"\"\n",
    "\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import scipy.stats as st\n",
    "\n",
    "data = pd.read_csv('data/data_w_mean_std.csv')\n",
    "\n",
    "parameter_grouped_data = dict(list(data.groupby('parameterid')))\n",
    "parameters = list(parameter_grouped_data.keys())\n",
    "\n",
    "outliers = pd.DataFrame(columns=data.columns)\n",
    "\n",
    "for parameter in parameters:\n",
    "    df = parameter_grouped_data[parameter]\n",
    "    mean=parameter_grouped_data[parameter]['mean'].mean()\n",
    "    standard_dev=parameter_grouped_data[parameter].standard_dev.mean()\n",
    "    usl = parameter_grouped_data[parameter].usl.mean()\n",
    "    lsl = parameter_grouped_data[parameter].lsl.mean()\n",
    "    outlier_df = df.loc[(df.mv>usl) | (df.mv<lsl)]\n",
    "    outliers=pd.concat([outliers,outlier_df],ignore_index=True)\n",
    "    \n",
    "outliers.to_csv('data/outlier_data.csv',index=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>lotid</th>\n",
       "      <th>deviceid</th>\n",
       "      <th>processid</th>\n",
       "      <th>operationid</th>\n",
       "      <th>parameterid</th>\n",
       "      <th>mv</th>\n",
       "      <th>wafer</th>\n",
       "      <th>usl</th>\n",
       "      <th>lsl</th>\n",
       "      <th>ucl</th>\n",
       "      <th>lcl</th>\n",
       "      <th>mean</th>\n",
       "      <th>standard_dev</th>\n",
       "      <th>target_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>F19320003.F1</td>\n",
       "      <td>SCL19MP0021_Process_R4</td>\n",
       "      <td>Product Flow 1.8V/3.3V w</td>\n",
       "      <td>MDF_YEDR1.033</td>\n",
       "      <td>CP_Pad_NIT_0.2_Delta</td>\n",
       "      <td>352.0</td>\n",
       "      <td>1</td>\n",
       "      <td>300.0</td>\n",
       "      <td>-125.0</td>\n",
       "      <td>125.0</td>\n",
       "      <td>-125.0</td>\n",
       "      <td>59.825243</td>\n",
       "      <td>48.404227</td>\n",
       "      <td>59.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>F18190003.F1</td>\n",
       "      <td>SCL18MP0015</td>\n",
       "      <td>Product Flow 1.8V/3.3V w</td>\n",
       "      <td>MWE_YEDR1.041</td>\n",
       "      <td>CP_Pad_Scan_0.16</td>\n",
       "      <td>710.0</td>\n",
       "      <td>1</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>244.228288</td>\n",
       "      <td>207.268746</td>\n",
       "      <td>244.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>F18190003.F1</td>\n",
       "      <td>SCL18MP0015</td>\n",
       "      <td>Product Flow 1.8V/3.3V w</td>\n",
       "      <td>MWE_YEDR1.041</td>\n",
       "      <td>CP_Pad_Scan_0.16</td>\n",
       "      <td>686.0</td>\n",
       "      <td>2</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>244.228288</td>\n",
       "      <td>207.268746</td>\n",
       "      <td>244.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          lotid                deviceid                 processid  \\\n",
       "0  F19320003.F1  SCL19MP0021_Process_R4  Product Flow 1.8V/3.3V w   \n",
       "1  F18190003.F1             SCL18MP0015  Product Flow 1.8V/3.3V w   \n",
       "2  F18190003.F1             SCL18MP0015  Product Flow 1.8V/3.3V w   \n",
       "\n",
       "     operationid           parameterid     mv  wafer    usl    lsl    ucl  \\\n",
       "0  MDF_YEDR1.033  CP_Pad_NIT_0.2_Delta  352.0      1  300.0 -125.0  125.0   \n",
       "1  MWE_YEDR1.041      CP_Pad_Scan_0.16  710.0      1  400.0    0.0  400.0   \n",
       "2  MWE_YEDR1.041      CP_Pad_Scan_0.16  686.0      2  400.0    0.0  400.0   \n",
       "\n",
       "     lcl        mean  standard_dev  target_value  \n",
       "0 -125.0   59.825243     48.404227          59.0  \n",
       "1    0.0  244.228288    207.268746         244.0  \n",
       "2    0.0  244.228288    207.268746         244.0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#-------displaying few above rows of the created dataset with outlier values-----------\n",
    "import pandas as pd\n",
    "data = pd.read_csv('data/outlier_data.csv')\n",
    "data.head(3)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2><u>Future Goals</u></h2>\n",
    "<p>Due to a very small amount of \"VALID\" data for each parameter the future activites depend on the availablity of better amounts of data with correct information implanted. However the above piece of the report would go same once a cleaner version of data is available.<p>\n",
    "<p>It would be essential to again list down the activities we seek for the future provided we get the valid forms of the required data\n",
    "    <ul>\n",
    "        <li> Removing the outliers. normalising the data and other <b>preprocessing</b> techniques to clean data. Finally, learning the effect of input parameters on the ouput.\n",
    "    <li> From the weights and relation obtained varying the input parameters through a range of LSL to USL and <b>find its value for optimal output</b> (best yield)</li>\n",
    "        <li><b>output the final values</b> that should be applied for the process.</li>\n",
    "     </ul>\n",
    "</p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2><u>References</u></h2>\n",
    "<p>\n",
    "[1] https://www.greycampus.com/blog/quality-management/how-to-measure-process-capability-and-process-performance\n",
    "<br>\n",
    "<br>\n",
    "[2] https://www.spcforexcel.com/knowledge/process-capability/cpk-alone-not-sufficient\n",
    "<br>\n",
    "<br>\n",
    "[3] https://www.isixsigma.com/tools-templates/capability-indices-process-capability/cp-cpk-pp-and-ppk-know-how-andwhen-use-them/\n",
    "<br>\n",
    "<br>\n",
    "[4] https://blog.minitab.com/blog/statistics-and-quality-data-analysis/cp-and-cpk-two-process-perspectives-one-processreality\n",
    "</p>"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}