Automated process for generating natural language descriptions of raster-based weather visualizations for output in written and audible form
View Patent ↗Generating specific and contextualized natural language descriptions based upon raster-based weather visualizations for a defined geographic region. The generated natural language descriptions are provided in a written and/or audible form. In some cases, these natural language descriptions are generated based on weather forecast data sets that indicate a relative motion of certain weather-related events.
1. A method comprising:
receiving a first weather data set including information indicative of: (i) a plurality of observed weather data points representing recent weather conditions, with the plurality of observed weather data points being a time ordered set of inputs, and (ii) a defined geographic region;
grouping a first portion of data points of the plurality of observed weather data points based, at least in part, upon the first portion of data points exceeding a first threshold;
creating a weather model with a deep neural network (DNN) by using the first portion of data points in the defined geographic region, with the weather model having information indicative of a relative motion of the recent weather conditions and a relative change of values of the first portion of data points over a period of time;
analyzing, by the machine learning module, the relative motion of the recent weather conditions and the relative change of values utilizing a locations database of the defined geographical region;
responsive to the analysis, creating, by a text description module, a natural language summarization of the recent weather conditions in the defined geographic region; and
responsive to the creation of the natural language summarization, generating a raster-based weather visualization having a defined length, the defined length based on an audible output length corresponding to a user-defined word rate.
2. The method of claim 1 wherein the natural language summarization is configured to be modifiable by a user based upon user adjustable thresholds.
3. The method of claim 1 wherein the natural language summarization includes an audible output.
4. The method of claim 1 further including:
adjusting, by a text-to-speech module, the audible output length of an audible output.
5. The method of claim 1 wherein a raster-based weather visualization includes information graphically displaying the relative change of values made to the received time ordered set of inputs.
6. A computer program product (CPP) comprising:
a machine-readable storage device; and
computer code stored on the machine-readable storage device, with the computer code including instructions and data for causing a processor(s) set to perform operations including the following:
receiving a first weather data set including information indicative of: (i) a plurality of observed weather data points representing recent weather conditions, with the plurality of observed weather data points being a time ordered set of inputs, and (ii) a defined geographic region;
grouping a first portion of data points of the plurality of observed weather data points based, at least in part, upon the first portion of data points exceeding a first threshold;
creating a weather model with a deep neural network (DNN) by using the first portion of data points in the defined geographic region, with the weather model having information indicative of a relative motion of the recent weather conditions and a relative change of values of the first portion of data points over a period of time;
analyzing, by the machine learning module, the relative motion of the recent weather conditions and the relative change of values utilizing a locations database of the defined geographical region;
responsive to the analysis, creating, by a text description module, a natural language summarization of the recent weather conditions in the defined geographic region; and
responsive to the creation of the natural language summarization, generating a raster-based weather visualization having a defined length, the defined length based on an audible output length corresponding to a user-defined word rate.
7. The CPP of claim 6 wherein the natural language summarization is configured to be modifiable by a user based upon user adjustable thresholds.
8. The CPP of claim 6 wherein the natural language summarization includes an audible output.
9. The CPP of claim 6 further including:
adjusting, by a text-to-speech module, the audible output length of an audible output.
10. The CPP of claim 6 wherein a raster-based weather visualization includes information graphically displaying the relative change of values made to the received time ordered set of inputs.
11. A computer system (CS) comprising:
a processor(s) set;
a machine-readable storage device; and
computer code stored on the machine-readable storage device, with the computer code including instructions and data for causing the processor(s) set to perform operations including the following:
receiving a first weather data set including information indicative of: (i) a plurality of observed weather data points representing recent weather conditions, with the plurality of observed weather data points being a time ordered set of inputs, and (ii) a defined geographic region;
grouping a first portion of data points of the plurality of observed weather data points based, at least in part, upon the first portion of data points exceeding a first threshold;
creating a weather model with a deep neural network (DNN) by using the first portion of data points in the defined geographic region, with the weather model having information indicative of a relative motion of the recent weather conditions and a relative change of values of the first portion of data points over a period of time;
analyzing, by the machine learning module, the relative motion of the recent weather conditions and the relative change of values utilizing a locations database of the defined geographical region;
responsive to the analysis, creating, by a text description module, a natural language summarization of the recent weather conditions in the defined geographic region; and
responsive to the creation of the natural language summarization, generating a raster-based weather visualization having a defined length, the defined length based on an audible output length corresponding to a user-defined word rate.
12. The CS of claim 11 wherein the natural language summarization is configured to be modifiable by a user based upon user adjustable thresholds.
13. The CS of claim 11 wherein the natural language summarization includes an audible output.
14. The CS of claim 11 further including:
adjusting, by a text-to-speech module, the audible output length of an audible output.
15. The CS of claim 11 wherein a raster-based weather visualization includes information graphically displaying the relative change of values made to the received time ordered set of inputs.