Load nino dataset

loadnino.py

# Data transformation from previous notebook
# col names in tao-all2.col from website
names = '''obs
year
month
day
date
latitude
longitude
zon.winds
mer.winds
humidity
air temp.
s.s.temp.'''.split('\n')

nino = pd.read_csv('data/tao-all2.dat.gz', sep=' ', names=names, na_values='.', 
                   parse_dates=[[1,2,3]])
nino.columns = [x.replace('.', '_').replace(' ', '_') for x in nino.columns]
nino['air_temp_F'] = nino.air_temp_ * 9/5 + 32
wind_cols = [x for x in nino.columns if x.endswith('winds')]
for c in wind_cols:
    nino['{}_mph'.format(c)] = nino[c] * 2.237
pd.to_datetime(nino.date, format='%y%m%d')
nino = nino.drop('obs', axis=1)

Here is a more modern version

zonal winds (west<0, east>0), meridional winds (south<0, north>0),

# Data transformation from previous notebook
# col names in tao-all2.col from website
names = '''obs
year
month
day
date
latitude
longitude
zon.winds
mer.winds
humidity
air temp.
s.s.temp.'''.split('\n')

nino = pd.read_csv('data/tao-all2.dat.gz', sep=' ', names=names, na_values='.', 
                   parse_dates=[[1,2,3]])

def clean_cols(val):
    return val.replace('.', '_').replace(' ', '_')

nino = (nino
  .rename(columns=clean_cols)
  .assign(air_temp_F=lambda df_: df_.air_temp_ * 9/5 + 32,
        zon_winds_mph=lambda df_: df_.zon_winds*2.237,
        mer_winds_mph=lambda df_: df_.mer_winds*2.237)
  .drop(columns='obs')
)