adding code for KPI calculations

Percent+of+Preventative+Maintenance+out+of+Total+Work+Orders+.py

@@ -0,0 +1,116 @@
+
+# coding: utf-8
+
+# In[1]:
+
+get_ipython().system(' pip install ipywidgets')
+
+
+# In[2]:
+
+import os
+import pandas as pd
+from pandas import ExcelWriter
+from pandas import ExcelFile
+import numpy as np
+
+
+# In[3]:
+
+#import excel sheet with work order data
+df =pd.read_excel('C:\\Users\\melanie.shimano\\Documents\\Copy15 of FY18 Outcome Budgeting--Data Validation.xlsx', sheetname ='FMD.FY16')
+
+
+# In[4]:
+
+#index the dataframe
+df.set_index("WO_ID",  inplace=True)
+
+
+# In[5]:
+
+#***SET THE FISCAL YEAR for analysis
+FYyear = 2015
+
+
+# In[6]:
+
+df.head()
+
+
+# In[7]:
+
+#remove rows where work order duration = open because won't be able to calculate anything from these
+#keep only work orders in FY for financial analysis indicated above in FYyear
+#keep only PM:CM relevant work orders: HVAC PM, boiler, chiller, HVAC, HVAC Infrastructure, HVAC repair
+df =df[df.WO_duration != 'open']
+df = df[df.FY == FYyear]
+
+
+# In[8]:
+
+#drop unneccesary columns in dataframe for ease of view
+col_list = ['BLDG#','FY','Month','prob_type']
+df = df[col_list]
+
+
+# In[9]:
+
+#count number of work orders by problem type
+df['count'] = df.groupby(['prob_type'])['prob_type'].transform('count')
+
+
+# In[10]:
+
+#keep only one of each problem type for analysis 
+df_prob_type = df.drop_duplicates(subset ='prob_type', keep='last')
+
+
+# In[11]:
+
+#drop values that are NaN
+df_prob_type.dropna()
+
+
+# In[12]:
+
+#keep only preventative and corrective HVAC related work order problem types
+df_pmcm = df_prob_type[df_prob_type['prob_type'].str.contains('HVAC|BOILER|CHILLERS|PREVENTIVE MAINT')==True]
+
+
+# In[13]:
+
+#create new dataframe of only preventative maintenance to then sum count
+df_pm = df_pmcm[df_pmcm['prob_type'].str.contains('PM|PREVENTIVE MAINT')==True]
+
+
+# In[14]:
+
+#total number of preventative maintenance work orders
+PM = df_pm['count'].sum()
+
+
+# In[15]:
+
+#sum the corrective maintenance tasks
+#create new dataframe of only corrective maintenance to then sum count
+#need to delete HVAC|PM counts that are included in the "HVAC" selection
+df_cm = df_pmcm[(df_pmcm['prob_type'].str.contains('HVAC|BOILER|CHILLERS')==True)&(~df_pmcm['prob_type'].str.contains('PM')==True)]
+
+
+# In[16]:
+
+#total number of correective maintenance work orders
+CM = df_cm['count'].sum()
+PMCM = (PM/CM)*100
+
+
+# In[17]:
+
+print('In Fiscal Year '+ str(FYyear) + ' the ratio of preventative maintenance versus corrective maintenance was ' + str(round (PM/CM,2)) +', or ' + str(round (PMCM,2)) + '%')
+
+
+# In[ ]:
+
+
+

Percent+of+Work+Orders+Closed+on+Time+KPI+Measure-Cleaned.py

@@ -0,0 +1,156 @@
+
+# coding: utf-8
+
+# In[1]:
+
+get_ipython().system(' pip install ipywidgets')
+
+
+# In[2]:
+
+import os
+import pandas as pd
+from pandas import ExcelWriter
+from pandas import ExcelFile
+import numpy as np
+import matplotlib.pyplot as plt
+from ipywidgets import *
+get_ipython().magic('matplotlib inline')
+
+
+# In[3]:
+
+#import excel sheet with work order data
+df =pd.read_excel('C:\\Users\\melanie.shimano\\Documents\\Copy15 of FY18 Outcome Budgeting--Data Validation.xlsx', sheetname ='FMD.FY16')
+
+
+# In[4]:
+
+#index the dataframe
+df.set_index("WO_ID",  inplace=True)
+
+
+# In[5]:
+
+#***SET THE FISCAL YEAR for analysis
+FYyear = 2016
+
+
+# In[6]:
+
+df.head()
+
+
+# In[7]:
+
+#remove rows where work order duration = open because won't be able to calculate anything from these
+#keep only work orders in FY for financial analysis indicated above in FYyear
+df =df[df.WO_duration != 'open']
+df = df[df.FY == FYyear]
+
+
+# In[8]:
+
+#add count of each prob_type to the data index
+df['Counts']=df.groupby(['prob_type'])['WO_duration'].transform('count')
+
+
+# In[9]:
+
+#drop unneccesary columns in dataframe for ease of view
+col_list = ['BLDG#','FY','Month','prob_type', 'WO_duration', 'Counts']
+df = df[col_list]
+
+
+# In[10]:
+
+df['total_duration']= df.groupby(['prob_type'])['WO_duration'].transform('sum')
+
+
+# In[11]:
+
+#calculate average work order duration per problem type (=counts/total_duration) and add value in new column
+df['avg_total_duration']=df['total_duration']/df['Counts']
+
+
+# In[12]:
+
+#check to see if individual work order closed at or equal to the average work order duration for that problem type
+#add new column "KPI_met" fill with "yes" if WO_duration <= avg_total_duration, otherwise fill 'no'
+df['closed_on_time']= np.where(df['WO_duration']<=df['avg_total_duration'], 'yes','no')
+
+
+# In[13]:
+
+#create new df of only KPIs that are met
+df_yes = df[df.closed_on_time == 'yes']
+
+
+# In[14]:
+
+#count number of KPI met per problem type
+df_count = df_yes.groupby('prob_type')['closed_on_time'].count()
+
+
+# In[15]:
+
+#turn series into a df to later merge with original dataframe and keep WO_ID
+df_KPI = pd.DataFrame({'prob_type': df_count.index, 'on_time_count':df_count.values})
+
+
+# In[16]:
+
+#new df (by merge) with number of KPI met in new column
+df_wo = df.assign(on_time_count =df['prob_type'].map(df_KPI.set_index('prob_type')['on_time_count']))
+
+
+# In[17]:
+
+#new df (by merge) with number of KPI met in new column
+df_wo = df.assign(on_time_count =df['prob_type'].map(df_KPI.set_index('prob_type')['on_time_count']))
+
+
+# In[18]:
+
+#calculate percentage of Work Orders that meet KPI and store in new column 'percent_KPI_met'
+df_wo['percent_KPI_met']=(df_wo['on_time_count']/df_wo['Counts'])*100
+
+
+# In[19]:
+
+#create dataframe of only unique problem types for high level analysis of KPIs met
+df_wo_final = df_wo.drop_duplicates(subset = 'prob_type', keep ='last')
+df_wo_final
+
+
+# In[20]:
+
+#get total values to calculate total number of KPI met
+WO_total_count = df_wo_final['Counts'].sum()
+
+WO_KPI_met_count = df_wo_final['on_time_count'].sum()
+
+percent_KPI_met = (WO_KPI_met_count/WO_total_count)*100
+
+
+# In[21]:
+
+#print total percentage of KPI met and total number of KPI met for the year
+print('Percent of Work Orders closed on time in FY '+ str(FYyear) +' = ' + str(round (percent_KPI_met, 2)) + "%, which means " +
+      str(round(WO_KPI_met_count,0)) + ' out of ' + str(round (WO_total_count,2)) + ' work orders were closed on time')
+
+
+# In[22]:
+
+#sort work order problem types by the percentage that did not meet the KPI (i.e. the percent that weren't closed on time)
+df_wo_final['not_on_time']=df_wo_final['Counts'] -df_wo_final['on_time_count']
+df_wo_final['percent_KPI_not_met']= (df_wo_final['not_on_time']/df_wo_final['Counts'])*100
+df_wo_final = df_wo_final.sort_values('percent_KPI_not_met')
+#graph the top 20 percent work orders by problem type that were not closed on time
+df_wo_final[50:].plot(x='prob_type', y='percent_KPI_not_met', kind='barh', title= 'Percent of Work Orders Not Closed on Time by Problem Type', figsize=(20,20))
+
+
+# In[ ]:
+
+
+