1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
def preprocessing(df: pd.DataFrame) -> pd.DataFrame:
# preprocessing
df['transaction_amount'] = df['transaction_amount'].astype(float)
df['transaction_date'] = pd.to_datetime(
df['transaction_date'], format="%Y-%m-%d %H:%M:%S").dt.date
df['transaction_ym'] = df['transaction_date'].apply(lambda x: x.strftime('%Y%m'))
df['transaction_year'] = df['transaction_year'].astype(object)
df['transaction_month'] = df['transaction_month'].astype(object)
df['year_of_construction'] = df['year_of_construction'].astype(int)
df['floor'] = df['floor'].astype(int)
df['dedicated_area'] = df['dedicated_area'].astype(float)
if 'monthly_rent' in df.columns:
df['monthly_rent'] = df['monthly_rent'].astype(int)
return df
def tran_cnt_per_month(df: pd.DataFrame) -> pd.DataFrame:
df = df.set_index('legal_dong')
print(len(df))
frames = []
print(f' dong counts : {len(df.index)}')
dong_names = df.index[:3]
for dong_name in tqdm.tqdm(dong_names):
data = df.loc[[dong_name]]
# data = applyParallel(
# data[['transaction_amount']].groupby(data.index), tempFunc)
data = data[['transaction_year']].groupby(data.index, 'transaction_year', 'transaction_month').mean()
data = data.reset_index()
frames.append(data)
result_df = pd.concat(frames, axis=0)
return result_df
result_df = df[['legal_dong', 'transaction_ym','transaction_amount']].groupby(['legal_dong','transaction_ym']).count()
result_df = result_df.rename(columns={'transaction_amount': 'count'})
result_df = pd.pivot(result_df.reset_index(), index='transaction_ym', columns='legal_dong', values='count')
result_df.plot(grid=True, figsize=(20, 10))