Pandasで日時データを要約する

Pythonで扱う日付と時刻

Max Shron

Data Scientist and Author

Pandasでデータを要約する

# ドック外にいた平均時間
rides['Duration'].mean()

Timedelta('0 days 00:19:38.931034482')

# ドック外にいた合計時間
rides['Duration'].sum()

Timedelta('3 days 22:58:10')

# ドック外にいた時間の割合
rides['Duration'].sum() / timedelta(days=91)

0.04348417785917786

# 各ステーションでの開始回数を数える
rides['Member type'].value_counts()

Member 236
Casual  54
Name: Member type, dtype: int64

# 会員種別ごとのライド割合
rides['Member type'].value_counts() / len(rides)

Member 0.814 
Casual 0.186 
Name: Member type, dtype: float64

# 期間（秒）列を追加
rides['Duration seconds'] = rides['Duration'].dt.total_seconds()

# 会員種別ごとの平均期間
rides.groupby('Member type')['Duration seconds'].mean()

Member type 
Casual 1994.667 
Member 992.280 
Name: Duration seconds, dtype: float64

# 月ごとの平均期間
rides.resample('ME', on = 'Start date')['Duration seconds'].mean()

Start date 
2017-10-31 1886.454 
2017-11-30 854.175 
2017-12-31 635.101 
Freq: ME, Name: Duration seconds, dtype: float64

# グループごとの件数
rides.groupby('Member type').size()

Member type
Casual     54
Member    236
dtype: int64

# グループごとの最初のライド
rides.groupby('Member type').first()

            Duration     ...     
Member type              ...
Casual      02:07:02     ...        
Member      00:03:01     ...

rides\
  .resample('ME', on = 'Start date')\
  ['Duration seconds']\
  .mean()\
  .plot()

rides\
  .resample('D', on = 'Start date')\
  ['Duration seconds']\
  .mean()\
  .plot()

Pythonで扱う日付と時刻