kaggle sharing-Electric Power Consumption

CY Wu

This is me who never ever stop learning.

kaggle sharing-Electric Power Consumption

Introduction

Tetouan是摩洛哥北部的一座城市，位於地中海沿岸，冬季氣候溫和多雨，夏季炎熱乾燥。摩洛哥在2015年關閉唯一煉油廠後，所有石油產品都是進口的(20萬桶/日)，耗電數據是從Amendis的監督控制和數據採集系統(SCADA)收集的，該系統是一家公共服務運營商，自2002年以來一直負責飲用水和電力的分配。配電網路的目的是為Tetouan的中低壓消費者提供服務。為此，Amendis確保從交付點到最終使用者(客戶)的電能交付和分配。分配的能源來自國家電力和飲用水辦公室。將高壓(63 kV)轉換為中壓(20 kV)后，可以傳輸和分配能量。[1]

數據包括 10 分鐘視窗內的 52,416 次能源消耗觀測值，觀測值包含:
1. Date time: 每十分鐘做紀錄
2. 溫度
3. 濕度
4. 風速
5. General Diffuse Flows(一般擴散流): 描述低溫（<0.2 °C 至 ~ 100 °C）流體，這些流體通過硫化物丘、裂縫熔岩流以及細菌墊和大型動物群的組合緩慢排出。
6. 擴散流
7. 區域1功耗
8. 區域2功耗
9. 區域3功耗

Part 1.: Humidy and Temperature w/ Poer Consumption plot[2]

Data

data在kaggle上Electric Power Consumption page上取得。

Importing package

Data cleaning & Check data

                    DatetimeIndex: 52416 entries, 2017-01-01 00:00:00 to 2017-12-30 23:50:00
                    Data columns (total 6 columns):
                        #   Column                  Non-Null Count  Dtype  
                    ---  ------                  --------------  -----  
                        0   temperature             52416 non-null  float64
                        1   humidity                52416 non-null  float64
                        2   windspeed               52416 non-null  float64
                        3   generaldiffuseflows     52416 non-null  float64
                        4   diffuseflows            52416 non-null  float64
                        5   powerconsumption_zone2  52416 non-null  float64
                    dtypes: float64(6)
                    memory usage: 2.8 MBRangeIndex: 1000 entries, 0 to 999

	temperature	humidity	windspeed	generaldiffuseflows	diffuseflows	powerconsumption_zone2
datetime
2017-01-01 00:00:00	6.559	73.8	0.083	0.051	0.119	16128.87538
2017-01-01 00:10:00	6.414	74.5	0.083	0.070	0.085	19375.07599
2017-01-01 00:20:00	6.313	74.5	0.080	0.062	0.100	19006.68693
2017-01-01 00:30:00	6.121	75.0	0.083	0.091	0.096	18361.09422
2017-01-01 00:40:00	5.921	75.7	0.081	0.048	0.085	17872.34043

Humidity and Temperature Scatter plot

Humidity and Temperature Scatter plot w/ Diffuse and Power Consumption

Humidity and Temperature Scatter plot w/ Diffuse and Power Consumption Summary

scatter-humidity & temperature saummry plot

Humidity and Temperature Scatter plot w/ Diffuse and Power Consumption (plotly.express)

All

First quantile

Second quantile

Third quantile

Fourth quantile

Part 2.: Correlation between Temperature and Power Consumption[3]

Combined plot-Powercoumption & Temperature-Date

Scatter plot between Powercoumption & Temperature for Linear and Ploynomial Regression

                    Function: power_consumption = temperature*220.41 + 16896.64
                    R^2 score: 0.17914777151851513

                    Function: power_consumption = 26861.28 * (temperature**0) + -899.67 * (temperature**1) + 29.16 * (temperature**2)
                    R^2 score: 0.26883141005067945

Conclusion

線性回歸沒有與原始資料相當擬合，因此我們嘗試用指數回歸去預測，得到的曲線與原始資料較相符，透過線性與指數回歸預測能得到溫度與電量使用量之間的R square值。

Source

[1] Electric Power Consumption
[2] Viz_Electricity
[3] Average Daily Power Consumption Prediction