Pandas Notebook 2, ATM350 Spring 2025

Pandas Notebook 2, ATM350 Spring 2025 #

Motivating Science Questions:#

What was the daily temperature and precipitation at Albany last year?
What were the the days with the most precipitation?

Motivating Technical Question:#

How can we use Pandas to do some basic statistical analyses of our data?

We’ll start by repeating some of the same steps we did in the first Pandas notebook.#

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

sns.set()

year = 2024

file = f'/spare11/atm350/common/data/climo_alb_{year}.csv'

Display the first five lines of this file using Python’s built-in readline function

fileObj = open(file)
nLines = 5
for n in range(nLines):
    line = fileObj.readline()
    print(line)

DATE,MAX,MIN,AVG,DEP,PCP,SNW,DPTH,HDD,CDD

2024-01-01,36,25,30.5,4.2,0.00,0.0,0,34,0

2024-01-02,38,24,31.0,4.9,0.00,0.0,0,34,0

2024-01-03,40,32,36.0,10.1,0.00,0.0,0,29,0

2024-01-04,43,19,31.0,5.3,T,T,0,34,0

df = pd.read_csv(file, dtype='string')

nRows = df.shape[0]
print (f"Number of rows = {nRows}" )
nCols = df.shape[1]
print (f"Number of columns = {nCols}")

date = df['DATE']
date = pd.to_datetime(date,format="%Y-%m-%d")

maxT = df['MAX'].astype("float32")
minT = df['MIN'].astype("float32")

Number of rows = 366
Number of columns = 10

Let’s generate the final timeseries we made in our first Pandas notebook, with all the “bells and whistles” included.#

from matplotlib.dates import DateFormatter, AutoDateLocator,HourLocator,DayLocator,MonthLocator

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, maxT, color='red',label = "Max T")
ax.plot (date, minT, color='blue', label = "Min T")
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Temperature (°F)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d967219400>

../../_images/4f69ad623b2556a2709edebe669127f814053ed4c6b7706d6432813f045d2725.png

Read in precip data. This will be more challenging due to the presence of T(races).#

Let’s remind ourselves what the Dataframe looks like, paying particular attention to the daily precip column (PCP).

df

	DATE	MAX	MIN	AVG	DEP	PCP	SNW	DPTH	HDD	CDD
0	2024-01-01	36	25	30.5	4.2	0.00	0.0	0	34	0
1	2024-01-02	38	24	31.0	4.9	0.00	0.0	0	34	0
2	2024-01-03	40	32	36.0	10.1	0.00	0.0	0	29	0
3	2024-01-04	43	19	31.0	5.3	T	T	0	34	0
4	2024-01-05	31	16	23.5	-2.0	T	T	0	41	0
...	...	...	...	...	...	...	...	...	...	...
361	2024-12-27	27	10	18.5	-8.9	0.03	0.3	2	46	0
362	2024-12-28	35	22	28.5	1.3	0.04	0.0	2	36	0
363	2024-12-29	57	33	45.0	18.1	0.04	0.0	1	20	0
364	2024-12-30	58	44	51.0	24.3	0.17	0.0	0	14	0
365	2024-12-31	51	30	40.5	14.0	T	0.0	0	24	0

366 rows × 10 columns

Exercise: define a Pandas DataSeries called precip and populate it with the requisite column from our Dataframe. Then print out its values.

TIP: After you have tried on your own, you can uncomment the first line of the cell below and re-run to load the solution.

# %load /spare11/atm350/common/mar06/02a.py
precip = df['PCP']
precip

    0.00
    0.00
    0.00
       T
       T
       ... 
  0.03
  0.04
  0.04
  0.17
     T
Name: PCP, Length: 366, dtype: string

The task now is to convert these values from strings to floating point values. Our task is more complicated due to the presence of strings that are clearly not numerical … such as “T” for trace.#

As we did in the first Pandas notebook with max temperatures greater than or equal to 90, create a subset of our `Dataframe` that consists only of those days where precip was a trace.#

traceDays = df[precip=='T']
traceDays

	DATE	MAX	MIN	AVG	DEP	PCP	SNW	DPTH	HDD	CDD
3	2024-01-04	43	19	31.0	5.3	T	T	0	34	0
4	2024-01-05	31	16	23.5	-2.0	T	T	0	41	0
10	2024-01-11	41	36	38.5	14.0	T	0.0	0	26	0
16	2024-01-17	22	9	15.5	-8.3	T	T	4	49	0
17	2024-01-18	30	10	20.0	-3.8	T	T	3	45	0
...	...	...	...	...	...	...	...	...	...	...
353	2024-12-19	40	27	33.5	4.0	T	0.0	0	31	0
357	2024-12-23	24	-5	9.5	-18.9	T	T	2	55	0
359	2024-12-25	27	8	17.5	-10.4	T	T	2	47	0
360	2024-12-26	20	5	12.5	-15.1	T	T	2	52	0
365	2024-12-31	51	30	40.5	14.0	T	0.0	0	24	0

78 rows × 10 columns

traceDays.shape

(78, 10)

traceDays.shape[0]

Exercise: print out the total # of days where a trace of precip was measured. Hint: look back at how we calculated the total # of 90 degree days in our first Pandas notebook ... we used the shape attribute.

# %load /spare11/atm350/common/mar06/02b.py
print (df[precip=='T'].shape)
numTraceDays = df[precip == 'T'].shape[0]
print (f"The total # of days in Albany in {year} that had a trace of precipitation was {numTraceDays}")

(78, 10)
The total # of days in Albany in 2024 that had a trace of precipitation was 78

Getting back to our task of converting precip amounts from strings to floating point numbers, one thing we could do is to create a new array and populate it via a loop, where we’d use an `if-else` logical test to check for Trace values and set the precip value to 0.00 for each day accordingly.#

There is a more efficient way to do this, though!#

See https://stackoverflow.com/questions/49154068/fixing-a-typeerror-when-using-pandas-after-replacing-a-string-with-a-floating-po?rq=1

We use the `loc` method of Pandas to find all elements of a DataSeries with a certain value, and then change that value to something else, all in the same line of code!#

In this case, let’s set all values of ‘T’ to ‘0.00’#

The line below is what we want! Before we execute it, let’s break it up into pieces.

df.loc[df['PCP'] =='T', ['PCP']] = '0.00'

First, create a `Series` of booleans corresponding to the specified condition.#

df['PCP'] == 'T'

    False
    False
    False
     True
     True
       ...  
  False
  False
  False
  False
   True
Name: PCP, Length: 366, dtype: boolean

Next, build on that cell by using `loc` to display all rows that correspond to the condition being True.#

df.loc[df['PCP'] == 'T']

	DATE	MAX	MIN	AVG	DEP	PCP	SNW	DPTH	HDD	CDD
3	2024-01-04	43	19	31.0	5.3	T	T	0	34	0
4	2024-01-05	31	16	23.5	-2.0	T	T	0	41	0
10	2024-01-11	41	36	38.5	14.0	T	0.0	0	26	0
16	2024-01-17	22	9	15.5	-8.3	T	T	4	49	0
17	2024-01-18	30	10	20.0	-3.8	T	T	3	45	0
...	...	...	...	...	...	...	...	...	...	...
353	2024-12-19	40	27	33.5	4.0	T	0.0	0	31	0
357	2024-12-23	24	-5	9.5	-18.9	T	T	2	55	0
359	2024-12-25	27	8	17.5	-10.4	T	T	2	47	0
360	2024-12-26	20	5	12.5	-15.1	T	T	2	52	0
365	2024-12-31	51	30	40.5	14.0	T	0.0	0	24	0

78 rows × 10 columns

Further build this line of code by only returning the column of interest.#

df.loc[df['PCP'] =='T', ['PCP']]

	PCP
3	T
4	T
10	T
16	T
17	T
...	...
353	T
357	T
359	T
360	T
365	T

78 rows × 1 columns

Finally, we have arrived at the full line of code! Take the column of interest, in this case precip only on those days where a trace was measured, and set its value to 0.00.#

df.loc[df['PCP'] =='T', ['PCP']] = '0.00'

df['PCP']

    0.00
    0.00
    0.00
    0.00
    0.00
       ... 
  0.03
  0.04
  0.04
  0.17
  0.00
Name: PCP, Length: 366, dtype: string

This operation actually modifies the `Dataframe` in place … i.e., the individual cell values have changed, and henceforth in the notebook, the `Dataframe` will reflect these changed values. We can prove this by printing out a row from a date that we know had a trace amount.#

But first, how do we simply print a specific row from a dataframe? Since we know that Jan. 4 had a trace of precip, try this:#

jan04 = df['DATE'] == '2024-01-04'
jan04

    False
    False
    False
     True
    False
       ...  
  False
  False
  False
  False
  False
Name: DATE, Length: 366, dtype: boolean

That produces a series of booleans; the one matching our condition is `True`. Now we can retrieve all the values for this date.#

df[jan04]

	DATE	MAX	MIN	AVG	DEP	PCP	SNW	DPTH	HDD	CDD
3	2024-01-04	43	19	31.0	5.3	0.00	T	0	34	0

We see that the precip has now been set to 0.00.#

Having done this check, and thus re-set the values, let’s now convert this series into floating point values.#

precip = df['PCP'].astype("float32")

precip

    0.00
    0.00
    0.00
    0.00
    0.00
       ... 
  0.03
  0.04
  0.04
  0.17
  0.00
Name: PCP, Length: 366, dtype: float32

Plot each day’s precip total.#

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, precip, color='blue', marker='+',label = "Precip")
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Precip (in.)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d966ff0500>

../../_images/2701c3b323b584848b5db8232a516266dac790472300624d609b3c5efdfaa4be.png

What if we just want to pick a certain time range? One simple way is to just pass in a subset of our x and y to the `plot` method.#

# Plot out just the trace for October. Corresponds to Julian days 215-246 ... thus, indices 214-245 (why?).
fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date[214:245], precip[214:245], color='blue', marker='+',label = "Precip")
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Precip (in.)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d9670afe60>

../../_images/bda3b596827db5084a2958d81264e27a83368d734b5f6cd931917df42b57db73.png

Exercise: print out a table of days with precip amounts of at least 1.00 inches. In a separate cell, print out the total # of such days.

# %load '/spare11/atm350/common/mar06/02c.py'
wetDays = df[precip>=1.00]
wetDays

	DATE	MAX	MIN	AVG	DEP	PCP	SNW	DPTH	HDD	CDD
8	2024-01-09	43	19	31.0	6.2	1.03	0.5	5	34	0
68	2024-03-09	45	36	40.5	7.3	1.05	T	0	24	0
82	2024-03-23	33	25	29.0	-9.1	1.69	4.4	2	36	0
125	2024-05-05	57	49	53.0	-3.2	1.24	0.0	0	12	0
221	2024-08-09	82	67	74.5	2.2	2.91	0.0	0	0	10
231	2024-08-19	82	59	70.5	-0.7	1.89	0.0	0	0	6
325	2024-11-21	51	40	45.5	7.0	1.17	0.0	0	19	0
332	2024-11-28	39	33	36.0	0.0	1.16	3.8	0	29	0
345	2024-12-11	52	36	44.0	12.2	1.39	T	0	21	0

# Split the cell here so the table above will be displayed!
numWetDays = wetDays.shape[0]
print (f"The total # of days in Albany in {year} that had at least 1.00 in. of precip was {numWetDays}" )

The total # of days in Albany in 2024 that had at least 1.00 in. of precip was 9

Pandas has a function to compute the cumulative sum of a series. We’ll use it to compute and graph Albany’s total precip over the year.#

precipTotal = precip.cumsum()

precipTotal

     0.000000
     0.000000
     0.000000
     0.000000
     0.000000
         ...    
  41.939995
  41.979996
  42.019997
  42.189995
  42.189995
Name: PCP, Length: 366, dtype: float32

We can see that the final total is in the last element of the precipTotal array. How can we explicitly print out just this value?#

One of the methods available to us in a Pandas DataSeries is `values`. Let’s display it:#

precipTotal.values

array([ 0.       ,  0.       ,  0.       ,  0.       ,  0.       ,
        0.15     ,  0.5      ,  0.5      ,  1.53     ,  2.       ,
        2.       ,  2.       ,  2.57     ,  2.58     ,  2.58     ,
        2.83     ,  2.83     ,  2.83     ,  2.83     ,  2.84     ,
        2.84     ,  2.84     ,  2.9399998,  3.4899998,  3.5599997,
        4.4599996,  4.47     ,  5.04     ,  5.1      ,  5.1      ,
        5.1      ,  5.1      ,  5.11     ,  5.11     ,  5.11     ,
        5.11     ,  5.11     ,  5.11     ,  5.11     ,  5.11     ,
        5.11     ,  5.11     ,  5.11     ,  5.11     ,  5.11     ,
        5.13     ,  5.1400003,  5.1500006,  5.1500006,  5.160001 ,
        5.160001 ,  5.160001 ,  5.160001 ,  5.270001 ,  5.270001 ,
        5.270001 ,  5.270001 ,  5.270001 ,  5.550001 ,  5.550001 ,
        5.550001 ,  6.000001 ,  6.000001 ,  6.000001 ,  6.710001 ,
        7.610001 ,  8.150002 ,  8.150002 ,  9.200002 ,  9.540002 ,
        9.540002 ,  9.540002 ,  9.540002 ,  9.540002 ,  9.620002 ,
        9.620002 ,  9.670002 ,  9.670002 ,  9.670002 ,  9.680002 ,
        9.680002 ,  9.820003 , 11.510002 , 11.510002 , 11.510002 ,
       11.510002 , 11.890002 , 11.890002 , 11.890002 , 11.890002 ,
       11.890002 , 11.890002 , 11.970002 , 12.450002 , 12.980001 ,
       12.980001 , 13.060001 , 13.060001 , 13.060001 , 13.060001 ,
       13.270001 , 13.700002 , 14.500002 , 14.540002 , 14.630002 ,
       14.630002 , 14.630002 , 14.780002 , 14.9400015, 14.9400015,
       14.9400015, 14.9400015, 14.9400015, 14.9400015, 14.9400015,
       14.9400015, 14.9400015, 14.9400015, 14.990002 , 15.000002 ,
       15.430002 , 15.430002 , 15.430002 , 15.430002 , 15.430002 ,
       16.670002 , 16.670002 , 16.670002 , 17.150002 , 17.150002 ,
       17.150002 , 17.150002 , 17.2      , 17.220001 , 17.220001 ,
       17.580002 , 17.580002 , 17.580002 , 17.580002 , 17.580002 ,
       17.580002 , 17.970001 , 17.970001 , 18.010002 , 18.010002 ,
       18.010002 , 18.010002 , 18.570002 , 18.570002 , 18.570002 ,
       18.570002 , 18.570002 , 18.570002 , 18.570002 , 18.570002 ,
       18.570002 , 18.570002 , 18.890001 , 19.400002 , 19.400002 ,
       20.130001 , 20.130001 , 20.130001 , 20.130001 , 20.130001 ,
       20.18     , 20.18     , 20.18     , 20.18     , 20.18     ,
       20.18     , 20.19     , 20.51     , 21.04     , 21.28     ,
       21.51     , 21.51     , 21.51     , 21.51     , 21.51     ,
       21.83     , 21.85     , 21.970001 , 21.970001 , 21.970001 ,
       21.970001 , 21.970001 , 22.130001 , 22.130001 , 22.130001 ,
       22.130001 , 22.130001 , 22.43     , 22.43     , 22.43     ,
       22.43     , 22.44     , 23.04     , 23.570002 , 23.570002 ,
       23.570002 , 23.570002 , 23.62     , 23.62     , 23.740002 ,
       23.760002 , 23.760002 , 23.760002 , 23.760002 , 23.760002 ,
       23.770002 , 23.770002 , 24.420002 , 24.420002 , 25.140001 ,
       25.62     , 25.730001 , 25.750002 , 26.380001 , 26.560001 ,
       27.2      , 30.11     , 30.11     , 30.18     , 30.18     ,
       30.18     , 30.18     , 30.18     , 30.18     , 30.18     ,
       30.460001 , 32.350002 , 32.38     , 32.38     , 32.38     ,
       32.38     , 32.38     , 32.4      , 32.4      , 32.4      ,
       32.4      , 32.4      , 32.4      , 32.550003 , 32.550003 ,
       32.550003 , 32.550003 , 32.550003 , 32.550003 , 32.550003 ,
       33.08     , 33.08     , 33.120003 , 33.120003 , 33.120003 ,
       33.120003 , 33.120003 , 33.120003 , 33.120003 , 33.120003 ,
       33.120003 , 33.120003 , 33.120003 , 33.120003 , 33.120003 ,
       33.120003 , 33.120003 , 33.120003 , 33.31     , 33.97     ,
       33.97     , 33.97     , 33.97     , 33.97     , 33.97     ,
       33.97     , 33.97     , 33.99     , 34.010002 , 34.010002 ,
       34.010002 , 34.030003 , 34.08     , 34.08     , 34.08     ,
       34.08     , 34.820004 , 35.210003 , 35.210003 , 35.210003 ,
       35.210003 , 35.210003 , 35.210003 , 35.210003 , 35.210003 ,
       35.210003 , 35.210003 , 35.210003 , 35.210003 , 35.210003 ,
       35.210003 , 35.210003 , 35.370003 , 35.370003 , 35.370003 ,
       35.390003 , 35.390003 , 35.390003 , 35.390003 , 35.390003 ,
       35.390003 , 35.390003 , 35.390003 , 35.390003 , 35.470005 ,
       35.490005 , 35.490005 , 35.490005 , 35.490005 , 35.490005 ,
       35.490005 , 35.490005 , 35.490005 , 35.490005 , 35.490005 ,
       36.660004 , 36.910004 , 37.050003 , 37.050003 , 37.050003 ,
       37.870003 , 37.870003 , 39.030003 , 39.050003 , 39.050003 ,
       39.050003 , 39.050003 , 39.050003 , 39.08     , 39.100002 ,
       39.100002 , 39.11     , 39.27     , 39.69     , 39.71     ,
       41.1      , 41.1      , 41.1      , 41.1      , 41.17     ,
       41.359997 , 41.389996 , 41.669994 , 41.669994 , 41.769993 ,
       41.849995 , 41.849995 , 41.849995 , 41.909996 , 41.909996 ,
       41.909996 , 41.939995 , 41.979996 , 42.019997 , 42.189995 ,
       42.189995 ], dtype=float32)

Exercise: It's an array! So, let's print out the last element of the array. What index # can we use?

# %load '/spare11/atm350/common/mar06/02d.py'
# Print out the value of the last element in the array
precipTotal.values[-1]

np.float32(42.189995)

Plot the timeseries of the cumulative precip for Albany over the year.#

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, precipTotal, color='blue', marker='.',label = "Precip")
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Precip (in.)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d966affe60>

../../_images/2d6a4eac927efeed1200261f0ed5f09670859c477ef1e94f3bb67e73552f3026.png

Pandas has a plethora of statistical analysis methods to apply on tabular data. An excellent summary method is `describe`.#

maxT.describe()

count    366.000000
mean      62.338799
std       19.533743
min       11.000000
25%       46.000000
50%       65.000000
75%       79.750000
max       96.000000
Name: MAX, dtype: float64

minT.describe()

count    366.000000
mean      42.967213
std       17.221647
min       -5.000000
25%       30.000000
50%       41.000000
75%       56.750000
max       76.000000
Name: MIN, dtype: float64

precip.describe()

count    366.000000
mean       0.115273
std        0.297395
min        0.000000
25%        0.000000
50%        0.000000
75%        0.050000
max        2.910000
Name: PCP, dtype: float64

Exercise: Why is the mean 0.12, but the median 0.00? Can you write a code cell that answers this question? Hint: determine how many days had a trace or less of precip.

First, express the condition where precip is equal to 0.00.
Then, determine the # of rows of that resulting series.

# %load /spare11/atm350/common/mar06/02e.py
subset = precip[precip == 0.00]
nRows = subset.shape[0]

print (f"The number of days where precip was a trace or less was {nRows}")
print ("Since this is represents more than half the days of the years, the median must = 0")

The number of days where precip was a trace or less was 236
Since this is represents more than half the days of the years, the median must = 0

We’ll wrap up by calculating and then plotting rolling means over a period of days in the year, in order to smooth out the day-to-day variations.#

First, let’s replot the max and min temperature trace for the entire year, day-by-day.

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, maxT, color='red',label = "Max T")
ax.plot (date, minT, color='blue', label = "Min T")
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Temperature (°F)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d966bb4500>

Now, let’s calculate and plot the daily mean temperature.

meanT = (maxT + minT) / 2.

meanT

    30.5
    31.0
    36.0
    31.0
    23.5
       ... 
  18.5
  28.5
  45.0
  51.0
  40.5
Length: 366, dtype: float32

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, meanT, color='green',label = "Mean T")
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Temperature (°F)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d9670382c0>

../../_images/927af964ed78a06ce13e5e369ce4b6ab951e7e049e56414e9ff3d35be99ec598.png

Next, let’s use Pandas’ rolling method to calculate the mean over a specified number of days. We’ll center the window at the midpoint of each period (thus, for a 30-day window, the first plotted point will be on Jan. 16 … covering the Jan. 1 –> Jan. 30 timeframe.

meanTr5 = meanT.rolling(window=5, center=True)
meanTr10 = meanT.rolling(window=10, center=True)
meanTr15 = meanT.rolling(window=15, center=True)
meanTr30 = meanT.rolling(window=30, center=True)

meanTr30.mean()

   NaN
   NaN
   NaN
   NaN
   NaN
       ..
 NaN
 NaN
 NaN
 NaN
 NaN
Length: 366, dtype: float64

meanTr5.mean()

     NaN
     NaN
    30.4
    30.4
    30.1
       ... 
  24.4
  31.1
  36.7
   NaN
   NaN
Length: 366, dtype: float64

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, meanT, color='green',label = "Mean T",alpha=0.2)
ax.plot (date, meanTr5.mean(), color='blue',label = "5 Day", alpha=0.3)
ax.plot (date, meanTr10.mean(), color='purple',label = "10 Day", alpha=0.3)
ax.plot (date, meanTr15.mean(), color='brown',label = "15 Day", alpha=0.3)
ax.plot (date, meanTr30.mean(), color='orange',label = "30 Day", alpha=1.0, linewidth=2)
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Temperature (°F)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d95e37fe60>

../../_images/593456021a723b3671e547b5c410f7ec058015c4ecfe6d6c92cecaf498ac7660.png

Display just the daily and 30-day running mean.

fig, ax = plt.subplots(figsize=(15,10))
ax.plot (date, meanT, color='green',label = "Mean T",alpha=0.2)
ax.plot (date, meanTr30.mean(), color='orange',label = "30 Day", alpha=1.0, linewidth=2)
ax.set_title (f"ALB Year {year}")
ax.set_xlabel('Date')
ax.set_ylabel('Temperature (°F)' )
ax.xaxis.set_major_locator(MonthLocator(interval=1))
dateFmt = DateFormatter('%b %d')
ax.xaxis.set_major_formatter(dateFmt)
ax.legend (loc="best")

<matplotlib.legend.Legend at 0x14d95e2b78f0>

../../_images/83b0d04e1341b786a5d62ac00c8a098833c53630832b5c297d8ceb258dd7e752.png