House vacancy analysis method and system for electric power data

Abstract

The invention discloses a housing vacancy analysis method and system based on electric power data, and the method comprises the steps: confirming a monitoring range, and obtaining the archive information of residential users with the standing time of more than one year, and taking the archive information as a housing vacancy monitoring range; data preparation: obtaining a daily electricity consumption curve of the to-be-monitored residential user in recent two years; constructing a model, analyzing power utilization index characteristics of vacant residential houses according to the electric quantity of the residential users, and constructing a residential power utilization characteristic analysis model; data analysis: according to the model, analyzing the electricity consumption characteristics of the residential building according to different spatial scales and spatial ranges, and dividing users into a high electricity consumption level, a normal electricity consumption level and a low electricity consumption level; and outputting a result, studying and judging the users in the range according to the users with the low power consumption level, obtaining the users with vacant houses, and calculating the vacancy rate of the houses. According to the invention, the housing vacancy condition can be accurately reflected, and the accuracy of the housing vacancy rate is improved.

Description

Technical Field

[0001] The present invention belongs to the field of power big data application and residential housing vacancy rate analysis, and specifically relates to a housing vacancy rate analysis method based on residential users' electricity consumption big data. Background Art

[0002] The vacancy situation in the property market is an important indicator reflecting the supply and demand situation in the market, and is also a "barometer" for judging whether there is a bubble. It has always attracted much attention from the market. Based on power data resources, State Grid Mengdong Power selected the electricity consumption data of all residential customers in the four leagues and cities in the jurisdiction in the past two years, drew on the experience of power data in the analysis of residential customer building vacancy, and organized an expert team to explore the application of power big data in property market control. Combined with data information such as residential customer electricity consumption, a building vacancy analysis model is constructed to analyze the overall characteristics of housing vacancy, urban and rural distribution characteristics, and detailed characteristics of prefecture-level cities and banners and counties, find out the current status of housing vacancy in the four leagues and cities, analyze housing utilization and change trends, and achieve preliminary identification and early warning of long-term vacant housing. Connect with relevant government departments such as prefecture-level cities, banners and counties to provide housing usage data to relevant government departments and serve the government in judging the supply situation in the property market. In the absence of official data, the existing housing vacancy rate statistics method, the data collected by private academic teams faces multiple considerations such as concepts, sample size and statistical methods, and often have large data errors. Therefore, a more reasonable and accurate housing vacancy analysis method is needed. Electricity data has the characteristics of wide coverage and high accuracy. It is closely related to the market and can effectively reflect the housing market conditions. Summary of the invention

[0003] The purpose of this invention is to solve the problem that in the absence of official data, the data compiled by private academic teams face multiple considerations such as concepts, sample size and statistical methods, and often result in large data errors.

[0004] In order to solve the above problems, the present invention provides a housing vacancy analysis method based on power data, comprising the following steps:

[0005] Step S1: confirm the monitoring scope and obtain the user file information of residents who have been registered for more than one year as the housing vacancy monitoring scope;

[0006] Step S1-1: Obtaining residential user profile information from the power system energy internet marketing service platform, limiting the user category to residential type, users who are not in the account cancellation status and whose account establishment date is one year from the current time;

[0007] Step S1-2: Match the electric energy meter number and power supply area information of the residential user; used to calculate the power consumption and count the residential vacancy situation by area;

[0008] Step S2: Data preparation, according to the range of residential users obtained in step S1, obtain the daily electricity consumption curve of the residential users to be monitored in the past two years;

[0009] Step S2-1: Obtain the electricity consumption information of residential users from the new generation electricity consumption information collection system and summarize it by day and month;

[0010] Step S2-2: Completing the missing daily electricity data by calculating the average of the daily electricity data of the two days before and after to complete the missing data;

[0011] Step S2-3: For data with negative values ​​or excessive daily power consumption, modify the data by calculating the average of the daily power consumption data of the two days before and after;

[0012] Step S3: constructing a model, analyzing the electricity consumption index characteristics of vacant residential houses according to the electricity consumption of residential users obtained in step S2, and constructing a residential electricity consumption characteristic analysis model;

[0013] Step S3-1: constructing a residential electricity consumption characteristic analysis model using a K-Means clustering algorithm;

[0014] Step S3-2: Initialize the centroid and randomly select 3 cluster centers (k 1 , k 2 , k 3 ), using their coordinates as the initial values ​​of the cluster centers;

[0015] Step S3-3: Calculate the distance between the sample and the cluster center. The user's daily power consumption data is one-dimensional data. The Manhattan distance is used to calculate the distance between each data point and the cluster center (k 1 , k 2 , k 3 ), that is, the absolute value of the value obtained by directly subtracting the sample values:

[0016]

[0017] In the formula, d represents the distance from a data point in the sample to the cluster center; x represents the daily electricity consumption of the residential user, as a data point in the sample; μ represents the centroid of the sample; n is the number of data points in the sample; i represents the i-th residential user;

[0018] Step S3-4: assign each sample to the cluster with the closest cluster center;

[0019] Step S3-5: Update the coordinates of the cluster centers. For each cluster, calculate the average values ​​of all samples in it and use them as the new cluster center coordinates.

[0020]

[0021] In the formula, k is the new cluster center, x represents the daily electricity consumption of residential users as a data point in the sample, and n is the number of data points in the sample;

[0022] Step S3-6: Repeat steps S3-3 to S3-5 until the cluster centers (k1, k2, k3) no longer change, that is, the center points and groups do not change after many iterations, that is, the center point of this calculation is the same as the center point of the previous calculation, and the clustering is completed;

[0023] Step S4: Data analysis: According to the model, the residential electricity consumption characteristics are analyzed according to different spatial scales and spatial ranges, and users are divided into three categories: high electricity consumption level, normal electricity consumption level, and low electricity consumption level;

[0024] Step S5: Output the result, obtain the users with low electricity consumption level according to step S4, and then analyze the users within the range to obtain the users with vacant houses, and calculate the housing vacancy rate;

[0025] Step S5-1: Calculate residential vacancy monitoring and analysis indicators based on the low-power users classified in step S4 as the vacant users analysis scope;

[0026] Step S5-2: Calculate the average daily power consumption (Q AVG ) and the standard deviation of daily electricity consumption (σ), the formula for average daily electricity consumption is:

[0027]

[0028] In the formula, QAVG is the average daily electricity consumption, Q is the daily electricity consumption, t is the statistical duration, and the formula for the standard deviation of daily electricity consumption is:

[0029]

[0030] Where, N is the statistical duration;

[0031] Step S5-3: Daily average power consumption Q AVG Residential users with a daily electricity consumption standard deviation of σ<δ are considered as vacant users;

[0032] Parameter setting: The thresholds ε and δ are set after comprehensive consideration of factors such as the city's economic level, electricity consumption data, and government demand. The recommended value range is 0<ε<=1, σ / 10<=δ<=σ / 5;

[0033] Step S5-4: Calculate the residential vacancy rate (T), the formula is:

[0034]

[0035] Wherein, T is the residential vacancy rate, S_empty is the number of vacant users, and S_total is the total number of users.

[0036] In the preferred method, the final clustering centers (k1, k2, k3) generated in step S3 are respectively used as the high electricity consumption level, normal electricity consumption level, and low electricity consumption level in step S4.

[0037] A housing vacancy analysis system based on power data, characterized by including: a data analysis module, a K-Means clustering algorithm module, and a vacancy calculation module. The specific steps include:

[0038] Step M1: Confirm the monitoring scope, and obtain the residential user file information with a household establishment time of more than one year as the housing vacancy monitoring scope.

[0039] Step M1-1: Obtain the residential user file information from the energy Internet marketing service platform of the power system, and limit the user category to residential type, non-cancelled status and users with a household establishment date one year from the current time.

[0040] Step M1-2: Match the electricity meter numbers and power supply area information of residential users.

[0041] Step M2: Data preparation. The data analysis module obtains the daily electricity consumption curves of the residents to be monitored in the past two years according to the scope of residential users obtained in step S1.

[0042] Step M2-1: Obtain the electricity consumption information of residential users from the new generation electricity information collection system and summarize it by day and month.

[0043] Step M2-2: Complete the missing daily electricity data, and complete the missing data by calculating the average value of the daily electricity data of the previous and next two days.

[0044] Step M2-3: Modify the negative and excessive daily electricity data by calculating the average value of the daily electricity data of the previous and next two days.

[0045] Step M3: Build a model. According to the electricity consumption of residential users obtained in step S2, analyze the electricity consumption index characteristics of residential housing vacancies and build a residential electricity consumption characteristic analysis model.

[0046] Step M3-1: Build a residential electricity consumption characteristic analysis model through the K-Means clustering algorithm module.

[0047] Step M3-2: Initialize the centroid, randomly select 3 clustering centers (k 1 , k 2 , k 3 ), and use their coordinates as the initial values of the clustering centers.

[0048] Step M3-3: Calculate the distance between the sample and the cluster center. The user's daily power consumption data is one-dimensional data. The Manhattan distance is used to calculate the distance between each data point and the cluster center (k 1 , k 2 , k 3 ), that is, the absolute value of the value obtained by directly subtracting the sample values:

[0049]

[0050] In the formula, d represents the distance from a data point in the sample to the cluster center; x represents the daily electricity consumption of the residential user, as a data point in the sample; μ represents the centroid of the sample; n is the number of data points in the sample; i represents the i-th residential user;

[0051] Step M3-4: Assign each sample to the cluster with the closest cluster center;

[0052] Step M3-5: Update the coordinates of the cluster centers. For each cluster, calculate the average values ​​of all samples in it and use them as the new cluster center coordinates.

[0053]

[0054] In the formula, k is the new cluster center, x represents the daily electricity consumption of residential users as a data point in the sample, and n is the number of data points in the sample;

[0055] Step M3-6: Repeat steps S3-3 to S3-5 until the cluster centers (k1, k2, k3) no longer change, that is, the center points and groups do not change after many iterations, that is, the center point of this calculation is the same as the center point of the previous calculation, and the clustering is completed;

[0056] Step M4: Data analysis: According to the model, the residential electricity consumption characteristics are analyzed according to different spatial scales and spatial ranges, and users are divided into three categories: high electricity consumption level, normal electricity consumption level, and low electricity consumption level;

[0057] Step M5: Output the result, obtain the users with low electricity consumption level according to step S4, and then analyze the users within the range, obtain the housing vacancy users through the vacancy calculation module, and calculate the housing vacancy rate;

[0058] Step M5-1: Calculate residential vacancy monitoring and analysis indicators based on the low-power users classified in step S4 as the vacant users analysis scope;

[0059] Step M5-2: Calculate the average daily power consumption (Q AVG ) and the standard deviation of daily electricity consumption (σ), the formula for average daily electricity consumption is:

[0060]

[0061] Wherein, QAVG is the daily average power consumption, Q is the daily power consumption, t is the statistical duration, and the formula for the standard deviation of daily power consumption is:

[0062]

[0063] Wherein, N is the statistical duration;

[0064] Step M5-3: Take the daily average power consumption Q AVG < ε kWh, and the residential users with the standard deviation of daily power consumption σ < δ are determined as vacant users;

[0065] Parameter setting: Set the thresholds ε and δ after comprehensive consideration according to factors such as the urban economic level, power consumption data, and government demands. It is recommended that the value range is 0 < ε <= 1, σ / 10 <= δ <= σ / 5;

[0066] Step M5-4: The vacant calculation module calculates the residential vacancy rate (T), and the formula is:

[0067]

[0068] Wherein, T is the residential vacancy rate, S_empty is the number of vacant users, and S_total is the total number of users.

[0069] The beneficial effects of the present invention are: accurately reflecting the housing vacancy situation and improving the accuracy of the housing vacancy rate. Brief Description of the Drawings

[0070] Figure 1 It is a schematic diagram of the working process of the present invention. Detailed Embodiment

[0071] According to Figure 1 , a housing vacancy analysis method based on power data includes the following steps:

[0072] Step S1: Confirm the monitoring scope, and obtain the file information of residential users whose household establishment time reaches more than one year as the housing vacancy monitoring scope;

[0073] Step S1-1: Obtain the file information of residential users from the energy Internet marketing service platform of the power system, and limit the user category to residential type, non-cancelled account status and users whose household establishment date is one year from the current time;

[0074] Step S1-2: Match the electricity meter numbers and power supply area information of residential users;

[0075] Step S2: Data preparation, according to the obtained residential user scope in step S1, obtain the daily power consumption curves of the residential users to be monitored in the past 2 years;

[0076] Step S2-1: Obtaining residential users' electricity consumption information from the new generation electricity consumption information collection system and summarizing it by day and month;

[0077] Step S2-2: Completing the missing daily electricity data by calculating the average of the daily electricity data of the two days before and after to complete the missing data;

[0078] Step S2-3: For data with negative values ​​or excessive daily power consumption, modify the data by calculating the average of the daily power consumption data of the two days before and after;

[0079] Step S3: constructing a model, analyzing the electricity consumption index characteristics of vacant residential houses according to the electricity consumption of residential users obtained in step S2, and constructing a residential electricity consumption characteristic analysis model;

[0080] Step S3-1: constructing a residential electricity consumption characteristic analysis model using a K-Means clustering algorithm;

[0081] Step S3-2: Initialize the centroid and randomly select 3 cluster centers (k 1 , k 2 , k 3 ), using their coordinates as the initial values ​​of the cluster centers;

[0082] Step S3-3: Calculate the distance between the sample and the cluster center. The user's daily power consumption data is one-dimensional data. The Manhattan distance is used to calculate the distance between each data point and the cluster center (k 1 , k 2 , k 3 ), that is, the absolute value of the value obtained by directly subtracting the sample values:

[0083]

[0084] In the formula, d represents the distance from a data point in the sample to the cluster center; x represents the daily electricity consumption of the residential user, as a data point in the sample; μ represents the centroid of the sample; n is the number of data points in the sample; i represents the i-th residential user;

[0085] Step S3-4: assign each sample to the cluster with the closest cluster center;

[0086] Step S3-5: Update the coordinates of the cluster centers. For each cluster, calculate the average values ​​of all samples in it and use them as the new cluster center coordinates.

[0087]

[0088] In the formula, k is the new cluster center, x represents the daily electricity consumption of residential users as a data point in the sample, and n is the number of data points in the sample;

[0089] Step S3-6: Repeat steps S3-3 to S3-5 until the cluster centers (k1, k2, k3) no longer change, that is, the center points and groups no longer change after multiple iterations, which means the center points calculated this time are the same as those of the previous time, and the clustering ends;

[0090] Step S4: Data analysis. According to the model, analyze the electricity consumption characteristics of residential houses according to different spatial scales and spatial ranges, and divide users into three categories: high electricity consumption level, normal electricity consumption level, and low electricity consumption level;

[0091] Step S5: Output results. Obtain the users with low electricity consumption levels according to step S4, then conduct research and judgment on the users within this range to obtain the vacant housing users, and calculate the housing vacancy rate;

[0092] Step S5-1: Based on the users with low electricity consumption levels divided in step S4, use them as the analysis range for vacant user research and judgment, and calculate the residential vacancy monitoring and analysis indicators;

[0093] Step S5-2: Calculate the average daily electricity consumption (Q AVG ) and the standard deviation of daily electricity consumption (σ) of the residential users within the range. The formula for the average daily electricity consumption is:

[0094]

[0095] In the formula, QAVG is the average daily electricity consumption, Q is the daily electricity consumption, t is the statistical duration. The formula for the standard deviation of daily electricity consumption is:

[0096]

[0097] In the formula, N is the statistical duration;

[0098] Step S5-3: Determine the residential users with an average daily electricity consumption Q AVG < ε kWh and a standard deviation of daily electricity consumption σ < δ as vacant users;

[0099] Parameter setting: Set the thresholds ε and δ after comprehensive consideration of factors such as the urban economic level, electricity consumption data, and government requirements. It is recommended that the value range be 0 < ε <= 1, σ / 10 <= δ <= σ / 5;

[0100] Step S5-4: Calculate the housing vacancy rate (T). The formula is:

[0101]

[0102] In the formula, T is the housing vacancy rate, Sempty is the number of vacant users, and Stotal is the total number of users.

[0103] The above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A housing vacancy analysis method based on power data, characterized in that: The following steps are involved: Step S1: confirm the monitoring scope and obtain the user file information of residents who have been registered for more than one year as the housing vacancy monitoring scope; Step S1-1: Obtaining residential user profile information from the power system energy internet marketing service platform, limiting the user category to residential type, users who are not in the account cancellation status and whose account establishment date is one year from the current time; Step S1-2: Matching the electric energy meter number and power supply area information of the residential user; Step S2: Data preparation, according to the range of residential users obtained in step S1, obtain the daily electricity consumption curve of the residential users to be monitored in the past two years; Step S2-1: Obtain the electricity consumption information of residential users from the new generation electricity consumption information collection system and summarize it by day and month; Step S2-2: Completing the missing daily electricity data by calculating the average of the daily electricity data of the two days before and after to complete the missing data; Step S2-3: For data with negative values ​​or excessive daily power consumption, modify the data by calculating the average of the daily power consumption data of the two days before and after; Step S3: constructing a model, analyzing the electricity consumption index characteristics of vacant residential houses according to the electricity consumption of residential users obtained in step S2, and constructing a residential electricity consumption characteristic analysis model; Step S3-1: constructing a residential electricity consumption characteristic analysis model using a K-Means clustering algorithm; Step S3-2: Initialize the centroid, randomly select three cluster centers (k1, k2, k3), and use their coordinates as the initial values ​​of the cluster centers; Step S3-3: Calculate the distance between the sample and the cluster center. The user's daily power consumption data is one-dimensional data. The distance between each data point and the cluster center (k1, k2, k3) is calculated by Manhattan distance, that is, the absolute value of the value obtained by directly subtracting the sample value is taken: In the formula, d represents the distance from a data point in the sample to the cluster center; x represents the daily electricity consumption of the residential user, as a data point in the sample; μ represents the centroid of the sample; n is the number of data points in the sample; i represents the i-th residential user; Step S3-4: assign each sample to the cluster with the closest cluster center; Step S3-5: Update the coordinates of the cluster centers. For each cluster, calculate the average values ​​of all samples in it and use them as the new cluster center coordinates. In the formula, k is the new cluster center, x represents the daily electricity consumption of residential users as a data point in the sample, and n is the number of data points in the sample; Step S3-6: Repeat steps S3-3 to S3-5 until the cluster centers (k1, k2, k3) no longer change, that is, the center points and groups do not change after many iterations, that is, the center point of this calculation is the same as the center point of the previous calculation, and the clustering is completed; Step S4: Data analysis: According to the model, the residential electricity consumption characteristics are analyzed according to different spatial scales and spatial ranges, and users are divided into three categories: high electricity consumption level, normal electricity consumption level, and low electricity consumption level; Step S5: Output the result, obtain the users with low electricity consumption level according to step S4, and then analyze the users within the range to obtain the users with vacant houses, and calculate the housing vacancy rate; Step S5-1: Take the users with low electricity consumption levels divided in Step S4 as the scope of analysis for vacant user judgment, and calculate the residential vacancy monitoring and analysis indicators; Step S5-2: Calculate the average daily power consumption (Q AVG ) and the standard deviation of daily electricity consumption (σ), the formula for average daily electricity consumption is: In the formula, QAVG is the average daily electricity consumption, Q is the daily electricity consumption, t is the statistical duration, and the formula for the standard deviation of daily electricity consumption is: In the formula, N is the statistical duration; Step S5-3: Daily average power consumption Q AVG Residential users with a daily electricity consumption standard deviation of σ<δ are considered as vacant users; Parameter setting: Set the thresholds ε and δ after comprehensive consideration of factors such as the urban economic level, electricity consumption data, and government requirements. It is recommended that the value range is 0 < ε <= 1, σ / 10 <= δ <= σ / 5; Step S5-4: Calculate the residential vacancy rate (T), and the formula is: In the formula, T is the residential vacancy rate, S_empty is the number of vacant users, and S_total is the total number of users.

2. The housing vacancy analysis method based on power data according to claim 1 is characterized in that: Take the final cluster centers (k1, k2, k3) generated in Step S3 as the high electricity consumption level, normal electricity consumption level, and low electricity consumption level in Step S4 respectively.

3. A housing vacancy analysis system based on power data, characterized in that: Including: Data analysis module, K-Means clustering algorithm module, vacancy calculation module. The specific steps include: Step M1: Confirm the monitoring scope, obtain the residential user file information with a household establishment time of more than one year as the housing vacancy monitoring scope; Step M1-1: Obtain the residential user file information from the energy Internet marketing service platform of the power system, and limit the user category to residential type, non-cancelled status and users with a household establishment date one year from the current time; Step M1-2: Match the electricity meter numbers and power supply area information of residential users; Step M2: Data preparation. The data analysis module obtains the daily electricity consumption curves of the residential users to be monitored in the past two years according to the residential user scope obtained in Step S1; Step M2-1: Obtain the electricity consumption information of residential users from the new generation of electricity information collection system and summarize it by day and month; Step M2-2: Complete the missing daily electricity data by calculating the average value of the daily electricity data of the previous and next two days; Step M2-3: Modify the negative and excessive daily electricity data by calculating the average value of the daily electricity data of the previous and next two days; Step M3: Build a model. Analyze the electricity consumption index characteristics of residential housing vacancy according to the electricity consumption of residential users obtained in Step S2, and build a residential electricity consumption characteristic analysis model; Step M3-1: Build a residential electricity consumption characteristic analysis model through the K-Means clustering algorithm module; Step M3-2: Initialize the centroid, randomly select 3 cluster centers (k1, k2, k3), and use their coordinates as the initial values of the cluster centers; Step M3-3: Calculate the distance between the sample and the cluster center. The user's daily electricity data belongs to one-dimensional data. Calculate the distance between each data point and the cluster centers (k1, k2, k3) through the Manhattan distance, that is, take the absolute value of the value obtained by directly subtracting the sample values: In the formula, d represents the distance from a data point in the sample to the cluster center; x represents the daily electricity consumption of residential users as a data point in the sample; μ represents the centroid of the sample; n is the number of data points in the sample; i represents the i-th residential user; Step M3-4: Assign each sample to the cluster where the cluster center closest to it is located; Step M3-5: Update the coordinates of the cluster centers. For each cluster, calculate the average value of all samples therein and use them as the coordinates of the new cluster centers. In the formula, k is the new cluster center, x represents the daily electricity consumption of residential users, which is a data point in the sample, and n is the number of data points in the sample. Step M3-6: Repeat Step S3-3 to Step S3-5 until the cluster centers (k1, k2, k3) no longer change, that is, the center points and groupings do not change after multiple iterations, which means the center points calculated this time are the same as those of the previous time, and the clustering ends. Step M4: Data analysis. According to the model, analyze the electricity consumption characteristics of residential houses at different spatial scales and spatial ranges, and classify users into three categories: high electricity consumption level, normal electricity consumption level, and low electricity consumption level. Step M5: Output the results. Obtain the users with low electricity consumption levels according to Step S4, and then conduct research and judgment on the users within this range. The housing vacancy users are obtained through the vacancy calculation module, and the housing vacancy rate is calculated. Step M5-1: Use the users with low electricity consumption levels divided in Step S4 as the research and analysis range for vacancy users, and calculate the residential vacancy monitoring and analysis indicators. Step M5-2: Calculate the average daily power consumption (Q AVG ) and the standard deviation of daily electricity consumption (σ), the formula for average daily electricity consumption is: In the formula, QAVG is the average daily electricity consumption, Q is the daily electricity consumption, t is the statistical duration, and the formula for the standard deviation of daily electricity consumption is: In the formula, N is the statistical duration. Step M5-3: Daily average power consumption Q AVG Residential users with a daily electricity consumption standard deviation of σ<δ are considered as vacant users; Parameter setting: Set the thresholds ε and δ after comprehensive consideration of factors such as the urban economic level, electricity consumption data, and government requirements. It is recommended that the value range is 0 < ε <= 1, σ / 10 <= δ <= σ / 5. Step M5-4: The vacancy calculation module calculates the residential vacancy rate (T), and the formula is: In the formula, T is the residential vacancy rate, S_empty is the number of vacancy users, and S_total is the total number of users.

Images