def test_mean_per_month():
df = get_dataframe("../../data/main/cleaned/mean/Amsterdam_o3_2015.csv")
measure_name = df.columns.values[-1]
expected_mean = df[df['DatetimeBegin'].str.contains(
'2015-01')][measure_name].sum() / (24 * 31)
df = mean_per_month(df)
assert expected_mean.round(2) == df[df[DATE_NAME] ==
first_day][measure_name].item()
def plot_mean_per_pol_city_year(city: CITY, pollutant: POLLUTANT, year: YEAR):
df = get_dataframe(
f'../../data/main/cleaned/mean/{city.name}_{pollutant.name}_{year.name}.csv'
)
df = mean_per_day(df)
sns.set_style("whitegrid", {'grid.linestyle': '-'})
plt.figure(figsize=(12, 6))
sns.lineplot(x="Date", y=f"mean {pollutant.name} (µg/m3)",
data=df).set_title(
f'{city.name} {pollutant.name} emissions {year.name}')
plt.show()
def test_get_mean_frame():
path = "../../data/main/cleaned/o3/Amsterdam/"
year = YEAR['2015']
df = concat_sets(path, year)
df = get_mean_frame(df, POLLUTANT.o3)
mean = df.iloc[0]['mean o3 (µg/m3)']
total = 0
nbr = 0
for path in Path(path).rglob('*' + year.name + '*.csv'):
df2 = get_dataframe(path)
total += df2[df2['DatetimeBegin'] ==
'2015-01-01 00:00:00 +01:00']["Concentration"].item()
nbr += 1
assert mean == (total / nbr)
def plot_mean_per_pol_city(city: CITY, pollutant: POLLUTANT, save=False):
df = pd.DataFrame()
for path in Path("../../data/main/cleaned/mean/").rglob(
f'{city.name}_{pollutant.name}_*.csv'):
print(path.absolute())
df_tmp = mean_per_month(get_dataframe(path))
df = pd.concat([df, df_tmp])
fig, axes = plt.subplots(figsize=(14, 6))
sns.set_style("whitegrid", {'grid.linestyle': '-'})
axes = sns.lineplot(x=DATE_NAME,
y=f'mean {pollutant.name} (µg/m3)',
data=df)
axes.set_title(f'{city.name} {pollutant.name} emissions')
# plt.xticks(rotation=40, ha='right')
plt.tight_layout()
plt.show()
if save:
axes.get_figure().savefig(f'{city.name}_plot_{pollutant.name}.png')
plt.close()
def test_get_dataframe():
df = emissions.get_dataframe("../../data/test/csv_test_1.csv")
assert len(df) == 3
assert len(df.columns) == 3
def test_get_dataframe_encoding():
df = emissions.get_dataframe("../../data/test/csv_test_2.csv",
'iso-8859-1')
assert len(df) == 3
assert len(df.columns) == 3
def get_stations_data(path: str) -> DataFrame:
df = emissions.get_dataframe(path, 'iso-8859-1')
return df