Land use planning recommendations using heterogeneous temporal datasets

1. A method for providing intelligent land use planning recommendations using heterogeneous temporal datasets in a computing environment by a processor, comprising:

receiving heterogeneous temporal datasets from a plurality of data sources, wherein at least a portion of the heterogeneous temporal datasets includes land use information of land regions, the land use information including environmental data, activity data of retail and commercial, services offered in the land regions, and traffic and route profile data servicing the land regions, and wherein at least another portion of the heterogeneous temporal datasets includes images of the land regions having encoded therein features representing characteristics of the land regions that change over time such that a recommended intervention of the land changes over time according to the characteristics of the land regions;

executing machine learning logic to generate a land use model trained in a first training stage using the heterogeneous temporal datasets, wherein training the land use model in the first training stage includes:

extracting, from each of a set of tuples of the heterogeneous temporal datasets comprising a geographical area tuple and a time interval tuple, geolocated heterogeneous chronological data related to a first time interval of a plurality of time intervals,

encoding the geolocated chronological data into an image matrix for the first time interval,

training the land use model using the image matrix to generate a classifier to associate a class to the image matrix, and

indexing both data of the images and associated text data describing land-use interventions in the selected geographical region, wherein an index of encoded geographical areas and interventions stores the data of the images and associated text data in a structured format comprising, for each geographical area tuple and time interval tuple, the image matrix, the class, a list of land-use interventions associated with the image matrix at the plurality of time intervals, and a corresponding polarity of a positive land-user intervention and a negative land-use intervention respectively assigned to each of the list of land-use interventions;

identifying, using the land use model, one or more geographical regions of the land regions having one or more similar characteristics or features to a selected geographical region, wherein the identifying includes matching features representing characteristics encoded within an image of the selected geographical region to the features representing the characteristics of at least one of the land regions by sorting the tuples of the geolocated heterogeneous chronological data to identify the one or more similar characteristics or features;

recommending one or more positive land-use interventions, one or more negative land-use interventions, or a combination thereof for the selected geographical region from the geolocated heterogeneous chronological data according to the generated land-use model, wherein the recommending includes suggesting the recommended intervention of the land of the selected geographical region according to the land use information of the one or more geographical regions during a given one of the plurality of time intervals in which the one or more similar characteristics or features are identified;

collecting feedback data relating to the one or more positive land-use interventions, the one or more negative land-use interventions, or the combination thereof; and

executing the machine learning logic to re-train, in a second training stage, the land use model according to the feedback data, wherein re-training the land use model in the second training stage utilizes the feedback data to iteratively enhance an accuracy of the recommendations of the one or more positive land-use interventions, the one or more negative land-use interventions, or the combination thereof.

2. The method of claim 1 , further including using the geolocated heterogeneous chronological data is used to identify the one or more similar characteristics or features, wherein the geolocated heterogeneous chronological data includes both structured data and unstructured data.

3. The method of claim 1 , further including learning the one or more positive land-use interventions, the one or more negative land-use interventions, or a combination thereof from one or more data sources during a selected time period.

4. The method of claim 1 , further including associating the one or more positive land-use interventions, the one or more negative land-use interventions, or a combination thereof to support the recommendation or non-support the recommendation.

5. The method of claim 1 , further including:

classifying the images using the machine learning logic, wherein the images are used as input when performing the training; and

recommending a list of positive land-use interventions and a list of negative land-use interventions of similar geographical areas and providing supporting evidences with the list of positive land-use interventions and the list of negative land-use interventions.

6. A system for providing intelligent land use planning recommendations using heterogeneous temporal datasets in a computing environment, comprising:

one or more computers with executable instructions that when executed cause the system to:

receive heterogeneous temporal datasets from a plurality of data sources, wherein at least a portion of the heterogeneous temporal datasets includes land use information of land regions, the land use information including environmental data, activity data of retail and commercial services offered in the land regions, and traffic and route profile data servicing the land regions, and wherein at least another portion of the heterogeneous temporal datasets includes images of the land regions having encoded therein features representing characteristics of the land regions that change over time such that a recommended intervention of the land changes over time according to the characteristics of the land regions;

execute machine learning logic to generate a land use model trained in a first training stage using the heterogeneous temporal datasets, wherein training the land use model in the first training stage includes:

extracting, from each of a set of tuples of the heterogeneous temporal datasets comprising a geographical area tuple and a time interval tuple, geolocated heterogeneous chronological data related to a first time interval of a plurality of time intervals,

encoding the geolocated chronological data into an image matrix for the first time interval,

training the land use model using the image matrix to generate a classifier to associate a class to the image matrix, and

indexing both data of the images and associated text data describing land-use interventions in the selected geographical region, wherein an index of encoded geographical areas and interventions stores the data of the images and associated text data in a structured format comprising, for each geographical area tuple and time interval tuple, the image matrix, the class, a list of land-use interventions associated with the image matrix at the plurality of time intervals, and a corresponding polarity of a positive land-user intervention and a negative land-use intervention respectively assigned to each of the list of land-use interventions;

identify, using the land use model, one or more geographical regions of the land regions having one or more similar characteristics or features to a selected geographical region, wherein the identifying includes matching features representing characteristics encoded within an image of the selected geographical region to the features representing the characteristics of at least one of the land regions by sorting the tuples of the geolocated heterogeneous chronological data to identify the one or more similar characteristics or features;

recommend one or more positive land-use interventions, one or more negative land-use interventions, or a combination thereof for the selected geographical region from the geolocated heterogeneous chronological data according to the generated land-use model, wherein the recommending includes suggesting the recommended intervention of the land of the selected geographical region according to the land use information of the one or more geographical regions during a given one of the plurality of time intervals in which the one or more similar characteristics or features are identified;

collect feedback data relating to the one or more positive land-use interventions, the one or more negative land-use interventions, or the combination thereof; and

execute the machine learning logic to re-train, in a second training stage, the land use model according to the feedback data, wherein re-training the land use model in the second training stage utilizes the feedback data to iteratively enhance an accuracy of the recommendations of the one or more positive land-use interventions, the one or more negative land-use interventions, or the combination thereof.

7. The system of claim 6 , wherein the executable instructions use the geolocated heterogeneous chronological data is used to identify the one or more similar characteristics or features, wherein the geolocated heterogeneous chronological data includes both structured data and unstructured data.

8. The system of claim 6 , wherein the executable instructions learn the one or more positive land-use interventions, the one or more negative land-use interventions, or a combination thereof from one or more data sources during a selected time period.

9. The system of claim 6 , wherein the executable instructions associate the one or more positive land-use interventions, the one or more negative land-use interventions, or a combination thereof to support the recommendation or non-support the recommendation.

10. The system of claim 6 , wherein the executable instructions:

classify the images using the machine learning logic, wherein the images are used as input when performing the training; and

recommend a list of positive land-use interventions and a list of negative land-use interventions of similar geographical areas and providing supporting evidences with the list of positive land-use interventions and the list of negative land-use interventions.

11. A computer program product for, by a processor, providing intelligent land use planning recommendations using heterogeneous temporal datasets, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

an executable portion that receives heterogeneous temporal datasets from a plurality of data sources, wherein at least a portion of the heterogeneous temporal datasets includes land use information of land regions, the land use information including environmental data, activity data of retail and commercial services offered in the land regions, and traffic and route profile data servicing the land regions, and wherein at least another portion of the heterogeneous temporal datasets includes images of the land regions having encoded therein features representing characteristics of the land regions that change over time such that a recommended intervention of the land changes over time according to the characteristics of the land regions;

an executable portion that executes machine learning logic to generate a land use model trained in a first training stage using the heterogeneous temporal datasets, wherein training the land use model in the first training stage includes:

extracting, from each of a set of tuples of the heterogeneous temporal datasets comprising a geographical area tuple and a time interval tuple, geolocated heterogeneous chronological data related to a first time interval of a plurality of time intervals,

encoding the geolocated chronological data into an image matrix for the first time interval,

training the land use model using the image matrix to generate a classifier to associate a class to the image matrix, and

indexing both data of the images and associated text data describing land-use interventions in the selected geographical region, wherein an index of encoded geographical areas and interventions stores the data of the images and associated text data in a structured format comprising, for each geographical area tuple and time interval tuple, the image matrix, the class, a list of land-use interventions associated with the image matrix at the plurality of time intervals, and a corresponding polarity of a positive land-user intervention and a negative land-use intervention respectively assigned to each of the list of land-use interventions;

an executable portion that identifies, using the land use model, one or more geographical regions of the land regions having one or more similar characteristics or features to a selected geographical region, wherein the identifying includes matching features representing characteristics encoded within an image of the selected geographical region to the features representing the characteristics of at least one of the land regions by sorting the tuples of the geolocated heterogeneous chronological data to identify the one or more similar characteristics or features;

an executable portion that recommends one or more positive land-use interventions, one or more negative land-use interventions, or a combination thereof for the selected geographical region from the geolocated heterogeneous chronological data according to the generated land-use model, wherein the recommending includes suggesting the recommended intervention of the land of the selected geographical region according to the land use information of the one or more geographical regions during a given one of the plurality of time intervals in which the one or more similar characteristics or features are identified;

an executable portion that collects feedback data relating to the one or more positive land-use interventions, the one or more negative land-use interventions, or the combination thereof; and

an executable portion that executes the machine learning logic to re-train, in a second training stage, the land use model according to the feedback data, wherein re-training the land use model in the second training stage utilizes the feedback data to iteratively enhance an accuracy of the recommendations of the one or more positive land-use interventions, the one or more negative land-use interventions, or the combination thereof.

12. The computer program product of claim 11 , further including an executable portion that uses the geolocated heterogeneous chronological data is used to identify the one or more similar characteristics or features, wherein the geolocated heterogeneous chronological data includes both structured data and unstructured data.

13. The computer program product of claim 11 , further including an executable portion that learns the one or more positive land-use interventions, the one or more negative land-use interventions, or a combination thereof from one or more data sources during a selected time period.

14. The computer program product of claim 11 , further including an executable portion that associates the one or more positive land-use interventions, the one or more negative land-use interventions, or a combination thereof to support the recommendation or non-support the recommendation.

15. The computer program product of claim 11 , further including an executable portion that:

classifies the images using the machine learning logic, wherein the images are used as input when performing the training; and

recommends a list of positive land-use interventions and a list of negative land-use interventions of similar geographical areas and providing supporting evidences with the list of positive land-use interventions and the list of negative land-use interventions.