notebook.community

Edit and run

Solar Panel Data Analysis

File Format





In [5]:



    


import gzip, shutil



    











In [6]:



    


# Compress CSV file for upload
with open('solar_data.csv', 'rb') as f_in:
    with gzip.open('solar_data.csv.gz', 'wb') as f_out:
        shutil.copyfileobj(f_in, f_out)



    











In [7]:



    


# Decompress CSV for processing
with gzip.open('solar_data.csv.gz', 'rb') as f_in:
    with open('solar_data.csv', 'wb') as f_out:
        shutil.copyfileobj(f_in, f_out)



    











In [8]:



    


with open('solar_data.csv') as f:
    for i in range(4):
        print(f.readline(), end='')
    for s in f:
        i += 1
print('••• ({:d} - 4 lines)\n{:s}'.format(i, s), end='')



    


















    






"Date & Time","Usage [kWh]","Generation [kWh]","Grid [kWh]","Solar [kWh]","Solar+ [kWh]"
2017-09-23 14:00:00,5398.379564444,5907.685191667,-556.027505000,5907.685191667,5954.407069444
2017-09-23 13:59:00,5398.332943333,5907.629288611,-556.018223056,5907.629288611,5954.351166389
2017-09-23 13:58:00,5398.286305000,5907.573029167,-556.008601944,5907.573029167,5954.294906944
••• (525928 - 4 lines)
2016-09-21 17:00:00,-283.494235278,-37.552713889,-284.370767222,-37.552713889,0.876531944

















In [2]:



    


import pandas



    











In [3]:



    


df = pandas.read_csv('solar_data.csv', parse_dates=['Date & Time'])



    











In [4]:



    


df.take([0,1,2,-1])



    


















    
Out[4]:











  
    

      
      Date & Time
      Usage [kWh]
      Generation [kWh]
      Grid [kWh]
      Solar [kWh]
      Solar+ [kWh]
    

  
  
    

      0
      2017-09-23 14:00:00
      5398.379564
      5907.685192
      -556.027505
      5907.685192
      5954.407069
    

    

      1
      2017-09-23 13:59:00
      5398.332943
      5907.629289
      -556.018223
      5907.629289
      5954.351166
    

    

      2
      2017-09-23 13:58:00
      5398.286305
      5907.573029
      -556.008602
      5907.573029
      5954.294907
    

    

      525927
      2016-09-21 17:00:00
      -283.494235
      -37.552714
      -284.370767
      -37.552714
      0.876532
    

  




















In [5]:



    


i = df['Generation [kWh]'] - df['Solar [kWh]']
i.max()



    


















    
Out[5]:








0.0

















In [6]:



    


i = df['Usage [kWh]'] - df['Solar+ [kWh]'] - df['Grid [kWh]']
i.max()



    


















    
Out[6]:








1.0014673534897156e-09

















In [7]:



    


i = df['Solar+ [kWh]'] - df['Solar [kWh]']
i.min(), i.mean(), i.max()



    


















    
Out[7]:








(38.429245833000003, 42.896849498967505, 46.721877778000817)

















In [8]:



    


import matplotlib.pyplot as plt
import matplotlib.dates as mdates

%matplotlib inline



    











In [9]:



    


df.set_index('Date & Time',inplace=True)
fig, ax = plt.subplots(figsize=(15,6))
df.plot(ax=ax)

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %y'))
ax.xaxis.set_minor_locator(mdates.WeekdayLocator())

Polar Plots





In [10]:



    


def day2polar(dt):
    t = dt.timetuple()
    i = (t.tm_hour*60 + t.tm_min)*60 + t.tm_sec
    return 2*pandas.np.pi * i / (24*3600)

def year2polar(dt):
    t = dt.timetuple()
    if t.tm_yday > 182:
        return 366 - t.tm_yday
    else:
        return t.tm_yday



    











In [11]:



    


pandas.np.pi, df.index[0].timetuple()



    


















    
Out[11]:








(3.141592653589793,
 time.struct_time(tm_year=2017, tm_mon=9, tm_mday=23, tm_hour=14, tm_min=0, tm_sec=0, tm_wday=5, tm_yday=266, tm_isdst=-1))

















In [12]:



    


#df['Time'] = df.index.time
df['Time [rad]'] = df.index.map(day2polar)



    











In [13]:



    


#p = df.plot.pie('Date & Time', 'Solar+ [kWh]', figsize=(12,6))
ax = plt.subplot(111, projection='polar')
i=2000
ax.plot(df['Time [rad]'][::i], df['Solar+ [kWh]'][::i], alpha=0.8)



    


















    
Out[13]:








[<matplotlib.lines.Line2D at 0x24998433cf8>]

Solar Generation and Usage





In [14]:



    


#i = df['Solar+ [kWh]'].rolling(2)
df['Solar'] = df['Solar [kWh]'].diff(-1)
df['Usage'] = df['Usage [kWh]'].diff(-1)
df['Loss'] = df['Solar+ [kWh]'].diff(-1) - df['Solar']



    











In [15]:



    


df2 = df.drop('Usage [kWh],Generation [kWh],Grid [kWh],Solar [kWh],Solar+ [kWh],Time [rad]'.split(','), axis=1)
df2.columns



    


















    
Out[15]:








Index(['Solar', 'Usage', 'Loss'], dtype='object')

















In [16]:



    


fig, ax = plt.subplots(figsize=(15,6))
df2.plot(ax=ax, ylim=(-0.05,0.25))

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %y'))
ax.xaxis.set_minor_locator(mdates.WeekdayLocator())
#ax.yaxis.set_data_interval(-0.5,0.5)



    


















    
























In [17]:



    


fig, ax = plt.subplots(figsize=(15,6))
df2.iloc[-10000:].plot(ax=ax)



    


















    
Out[17]:








<matplotlib.axes._subplots.AxesSubplot at 0x24996d88438>










    
























In [18]:



    


fig, ax = plt.subplots(figsize=(15,6))
df2.iloc[:10000].plot(ax=ax)
plt.gca().invert_xaxis()



    


















    
























In [19]:



    


df3 = df2['09/24/2016':'09/23/2017']
fig, ax = plt.subplots(figsize=(15,6))
df3['Solar'].plot(ax=ax)

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %y'))
ax.xaxis.set_minor_locator(mdates.WeekdayLocator())



    


















    
























In [50]:



    


fig, ax = plt.subplots(figsize=(15,6))
df4 = df3.groupby(pandas.TimeGrouper('D'))
df4['Solar'].mean().plot(ax=ax, label='Mean')
df4['Solar'].min().plot(ax=ax, label='Min')
df4['Solar'].max().plot(ax=ax, label='Max')

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %d'))
ax.xaxis.set_minor_locator(mdates.WeekdayLocator())
legend = ax.legend(shadow=True, fontsize='x-large')

Solar Generation over the day





In [20]:



    


fig, ax = plt.subplots(figsize=(15,6))
cm = plt.get_cmap('hsv')
for i, d in enumerate(pandas.date_range('09/24/2016', periods=52, f='W')):
    # grab series
    s = df3[d.strftime('%D')]['Solar']
    # convert datetime to time of day
    t = s.index - d
    # 96 days between summer solstice (June 20) and start date
    ax.plot(t.total_seconds()/3600, s.values, alpha=0.5, c=cm((i+44)/52.0 % 1))
    
plt.xlim(5,21)
plt.xticks(tuple(range(5,21+1,1)))
plt.xlabel('Hour in day')



    


















    
Out[20]:








<matplotlib.text.Text at 0x2498b29bcf8>










    
























In [21]:



    


dst16 = pandas.date_range('03/13/2016','11/6/2016', freq='D')
dst17 = pandas.date_range('03/12/2017','11/5/2017', freq='D')
dst16



    


















    
Out[21]:








DatetimeIndex(['2016-03-13', '2016-03-14', '2016-03-15', '2016-03-16',
               '2016-03-17', '2016-03-18', '2016-03-19', '2016-03-20',
               '2016-03-21', '2016-03-22',
               ...
               '2016-10-28', '2016-10-29', '2016-10-30', '2016-10-31',
               '2016-11-01', '2016-11-02', '2016-11-03', '2016-11-04',
               '2016-11-05', '2016-11-06'],
              dtype='datetime64[ns]', length=239, freq='D')

















In [133]:



    


def plotDailySolar(dataframe, frequency='D', colormap='hsv', k='Solar'):
    fig, ax = plt.subplots(figsize=(15,6))
    cm = plt.get_cmap(colormap)
    p = {'D':365,'W':52}
    offset = {'D':96,'W':44}
    alpha = {'D':0.2, 'W':0.5}
    for i, d in enumerate(pandas.date_range('09/24/2016', periods=p[frequency], f=frequency)):
        # grab series
        s = dataframe[d.strftime('%D')][k]
        # convert datetime to time of day
        t = s.index - d
        # adjust for daylight savings
        if d in dst16 or d in dst17:
            hr = 1
        else: # not in daylight savings
            hr = 0
        c = cm((i+offset[frequency])*1.0/p[frequency] % 1)
        ax.plot(t.total_seconds()/3600 - hr, s.values, alpha=alpha[frequency], c=c)

    plt.xlim(4,20)
    plt.xticks(tuple(range(4,20+1,1)))
    plt.xlabel('Hour in day')
    return ax
    
plotDailySolar(df3, 'W')



    


















    
Out[133]:








<matplotlib.axes._subplots.AxesSubplot at 0x2498e1ffd30>










    
























In [23]:



    


ax = plotDailySolar(df3)

Solar Generation over the year





In [130]:



    


fig, ax = plt.subplots(figsize=(15,6))
df4 = df3.groupby(pandas.TimeGrouper('D'))
df4['Solar [norm]'].mean().plot(ax=ax, label='Mean', c='#7fa7ff', linewidth=1)
df4['Solar [norm]'].max().plot(ax=ax, label='Max', c='#c771fc', linewidth=1)
solar_max = df4['Solar'].sum()
solar_max.apply(lambda x: x / solar_max.max()).plot(ax=ax, label='Daily', c='blue')
df4['Solar [norm]'].apply(lambda column: (column > 0).sum() / len(column)).plot(ax=ax, label='Daylight', linewidth=3, c='#333333')

plt.ylim(0,1)
ax.set_yticks(pandas.np.linspace(0,1,12+1), False)
ax.set_yticks(pandas.np.linspace(0,1,24+1),True)
ax.grid(which='major', axis='y')

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %d'))
ax.xaxis.set_minor_locator(mdates.WeekdayLocator())
legend = ax.legend(loc='lower right', shadow=True, fontsize='large')
Date June 20 cumulative Dec 20 cumulative delta
Night 04:20 04:20 11:31 11:31 07:11
Astronomical Twilight 01:48 06:08 01:08 12:39 00:40
Nautical Twilight 01:29 07:37 01:10 13:49 00:19
Civil Twilight 01:09 08:46 01:04 14:53 00:05
Daylight 15:14 24:00 09:07 24:00 -06:07 




In [204]:



    


def percent_daylight(t, twilight=''):
    """Given a pandas Timestamp, calculate the percentage of daylight in a 24 hour period.
Can optionally include the duration of "Civil", "Nautical", or "Astronomical" twilight."""
    solstice = pandas.Timestamp('2015-06-20')
    year = pandas.Timedelta('365 days 06:00:00')
    scale = 2 * pandas.np.pi
    day_fraction = lambda v: pandas.Timedelta(v+':00') / pandas.Timedelta('1 day')
    # HH:MM length of night at summer solstice
    d={'Civil':'01:09','Nautical':'01:29','Astronomical':'01:48'}
    base = {k:day_fraction(v) for k,v in d.items()}
    # HH:MM delta between solstices
    d={'Civil':'00:05','Nautical':'00:19','Astronomical':'00:40'}
    delta = {k:day_fraction(v) for k,v in d.items()}
    # Calculate time in year, 0 = winter solstice, 1 = summer solstice
    i = pandas.np.cos((t - solstice)/year * scale) * 0.5 + 0.5
    # Calculate daylight
    j = i * day_fraction('06:07') + day_fraction('09:07')
    if twilight:
        j += i * delta[twilight] + base[twilight]
    return j

for d in pandas.date_range('09/24/2016',periods=12,freq='M'):
    print(d, percent_daylight(d))



    


















    






2016-09-30 00:00:00 0.482380810096
2016-10-31 00:00:00 0.42230945294
2016-11-30 00:00:00 0.386519633875
2016-12-31 00:00:00 0.382623032289
2017-01-31 00:00:00 0.413347719806
2017-02-28 00:00:00 0.463922493999
2017-03-31 00:00:00 0.530857471249
2017-04-30 00:00:00 0.589586641131
2017-05-31 00:00:00 0.627619646791
2017-06-30 00:00:00 0.632649121575
2017-07-31 00:00:00 0.603605410746
2017-08-31 00:00:00 0.547814747926

















In [242]:



    


fig, ax = plt.subplots(figsize=(15,6))
df4 = df3.groupby(pandas.TimeGrouper('D'))
df4['Solar [norm]'].mean().plot(ax=ax, label='Mean', c='#7fa7ff', linewidth=1)
df4['Solar [norm]'].max().plot(ax=ax, label='Max', c='#c771fc', linewidth=1)
solar_max = df4['Solar'].sum()
solar_max.apply(lambda x: x / solar_max.max()).plot(ax=ax, label='Daily', c='blue')
x = df4['Solar [norm]'].apply(lambda column: (column > 0).sum() / len(column))
x.plot(ax=ax, label='Daylight', linewidth=3, c='#333333')
df4['Daylight [exp]'].mean().plot(ax=ax, label='Daylight [exp]', linewidth=3, c='orange')
(x / df3['Daylight [exp]']).dropna().plot(ax=ax, label='Clear Skies', marker='.', linewidth=0, c='orange')

plt.ylim(0,1)
ax.set_yticks(pandas.np.linspace(0,1,12+1), False)
ax.set_yticks(pandas.np.linspace(0,1,24+1),True)
ax.grid(which='major', axis='y')

ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %d'))
ax.xaxis.set_minor_locator(mdates.WeekdayLocator())
legend = ax.legend(loc='lower right', shadow=True, fontsize='large')

Electricity Usage over the day





In [140]:



    


ax = plotDailySolar(df3, 'D', k='Usage')
ax.set_title('Daily Usage')
ax.set_xticks(range(0,24+1))
ax.set_xlim(0,24)



    


















    
Out[140]:








(0, 24)










    
























In [144]:



    


ax = plotDailySolar(df3, 'D', k='Usage [norm]')
ax.set_title('Daily Usage as Fraction of Solar Generation')
ax.set_xticks(range(0,24+1))
ax.set_xlim(0,24)



    


















    
Out[144]:








(0, 24)










    
























In [149]:



    


fig, ax = plt.subplots(figsize=(15,6))
df4 = df3.groupby(pandas.TimeGrouper('15M'))
df4['Usage [norm]'].mean().plot(ax=ax, label='Mean', linewidth=1) #c='#7fa7ff', 
#df4['Solar [norm]'].max().plot(ax=ax, label='Max', c='#c771fc', linewidth=1)
#solar_max = df4['Solar'].sum()
#solar_max.apply(lambda x: x / solar_max.max()).plot(ax=ax, label='Daily', c='blue')
#df4['Solar [norm]'].apply(lambda column: (column > 0).sum() / len(column)).plot(ax=ax, label='Daylight', linewidth=3, c='#333333')

plt.ylim(0,1)
ax.set_yticks(pandas.np.linspace(0,1,12+1), False)
ax.set_yticks(pandas.np.linspace(0,1,24+1),True)
ax.grid(which='major', axis='y')

#ax.xaxis.set_major_locator(mdates.MinuteLocator())
#ax.xaxis.set_major_formatter(mdates.DateFormatter('%H'))
legend = ax.legend(loc='lower right', shadow=True, fontsize='large')



    


















    
























In [152]:



    


pandas.TimeGrouper??

Content source: jimaples/jimaples.github.io

Similar notebooks:

notebook.community | gallery | about