Systems and methods for sentiment analysis and categorization
Described herein are systems and techniques to facilitate efficient determination of operator sentiment based on unstructured textual data exchanged between computing devices via communications channels. Unstructured textual data may be preprocessed to remove extraneous data and prepare the textual content for input to a machine-learned model trained to determine one or more sentiment scores based on textual data. The output of the model may be used to determine sentiment data and/or trends and to determine one or more subsequent actions.
1 . A method, comprising:
receiving, by an instruction generation system, unstructured textual data and metadata associated with a text communications exchange between a first computing device and a second computing device;
preprocessing the unstructured textual data and metadata at the instruction generation system by:
determining, based at least in part on the metadata, one or more portions of the unstructured textual data originating with the first computing device; and
generating, based at least in part on the one or more portions of the unstructured textual data originating with the first computing device, machine-learned model input data by removing extraneous data from the one or more portions of the unstructured textual data originating with the first computing device and tokenizing the one or more portions of the unstructured textual data originating with the first computing device;
executing, at the instruction generation system, a machine-learned model using the machine-learned model input data as input to generate operator data as output; and
transmitting, from the instruction generation system to the second computing device, instructions to generate an interface at the second computing device, the interface comprising an interface element comprising a graphical representation of the operator data.
2 . The method of claim 1 , wherein the machine-learned model comprises one or more of a large language model or a convolutional neural network.
3 . The method of claim 1 , further comprising transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface based at least in part on the operator data.
4 . The method of claim 3 , wherein the second instructions comprise instructions to present one or more communications channels associated with one or more subsequent communications exchanges between the first computing device and the second computing device.
5 . The method of claim 1 , further comprising:
determining, by the instruction generation system, an operator data trend based at least in part on the operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface, the second interface element comprising a second graphical representation of the operator data trend.
6 . The method of claim 5 , further comprising transmitting, from the instruction generation system to the second computing device, second instructions to generate a third interface element at the interface based at least in part on the operator data trend, the third interface element comprising data representing one or more actions determined based at least in part on the operator data trend.
7 . The method of claim 1 , further comprising:
generating aggregated operator data based at least in part on the operator data and historical operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface based at least in part on the aggregated operator data.
8 . A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors executing on an instruction generation system configured in a computing environment, cause the one or more processors to perform operations comprising:
receiving, by the instruction generation system, unstructured textual data and metadata associated with a text communications exchange between a first computing device and a second computing device;
preprocessing the unstructured textual data and metadata at the instruction generation system by:
determining, based at least in part on the metadata, one or more portions of the unstructured textual data originating with the first computing device; and
generating, based at least in part on the one or more portions of the unstructured textual data originating with the first computing device, machine-learned model input data by removing extraneous data from the one or more portions of the unstructured textual data originating with the first computing device and tokenizing the one or more portions of the unstructured textual data originating with the first computing device;
executing, at the instruction generation system, a machine-learned model using the machine-learned model input data as input to generate operator data as output; and
transmitting, from the instruction generation system to the second computing device, instructions to generate an interface at the second computing device, the interface comprising an interface element comprising a graphical representation of the operator data.
9 . The non-transitory computer-readable medium of claim 8 , wherein the machine-learned model comprises one or more of a large language model or a convolutional neural network.
10 . The non-transitory computer-readable medium of claim 8 , wherein the operations further comprise transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface based at least in part on the operator data.
11 . The non-transitory computer-readable medium of claim 10 , wherein the second instructions comprise instructions to present one or more communications channels associated with one or more subsequent communications exchanges between the first computing device and the second computing device.
12 . The non-transitory computer-readable medium of claim 8 , wherein the operations further comprise:
determining, by the instruction generation system, an operator data trend based at least in part on the operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface, the second interface element comprising a second graphical representation of the operator data trend.
13 . The non-transitory computer-readable medium of claim 12 , wherein the operations further comprise transmitting, from the instruction generation system to the second computing device, second instructions to generate a third interface element at the interface based at least in part on the operator data trend, the third interface element comprising data representing one or more actions determined based at least in part on the operator data trend.
14 . The non-transitory computer-readable medium of claim 8 , wherein the operations further comprise:
generating aggregated operator data based at least in part on the operator data and historical operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface based at least in part on the aggregated operator data.
15 . An instruction generation system, comprising:
one or more processors executing in a computing environment; and
a non-transitory memory storing computer-executable instructions that, when executed, cause the one or more processors to perform operations comprising:
receiving unstructured textual data and metadata associated with a text communications exchange between a first computing device and a second computing device;
preprocessing the unstructured textual data and metadata at the instruction generation system by:
determining, based at least in part on the metadata, one or more portions of the unstructured textual data originating with the first computing device; and
generating, based at least in part on the one or more portions of the unstructured textual data originating with the first computing device, machine-learned model input data by removing extraneous data from the one or more portions of the unstructured textual data originating with the first computing device and tokenizing the one or more portions of the unstructured textual data originating with the first computing device;
executing a machine-learned model using the machine-learned model input data as input to generate operator data as output; and
transmitting, to the second computing device, instructions to generate an interface at the second computing device, the interface comprising an interface element comprising a graphical representation of the operator data.
16 . The instruction generation system of claim 15 , wherein the operations further comprise:
determining, by the instruction generation system, an operator data trend based at least in part on the operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface, the second interface element comprising a second graphical representation of the operator data trend.
17 . The instruction generation system of claim 16 , wherein the operations further comprise transmitting, from the instruction generation system to the second computing device, second instructions to generate a third interface element at the interface based at least in part on the operator data trend, the third interface element comprising data representing one or more actions determined based at least in part on the operator data trend.
18 . The instruction generation system of claim 15 , wherein the operations further comprise:
generating aggregated operator data based at least in part on the operator data and historical operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface based at least in part on the aggregated operator data.
19 . The instruction generation system of claim 15 , wherein the operations further comprise:
determining one or more actions based at least in part on the operator data; and
transmitting, from the instruction generation system to the second computing device, second instructions to generate a second interface element at the interface, the second interface element comprising data representing the one or more actions.
20 . An instruction generation system, comprising:
means for receiving unstructured textual data and metadata associated with a text communications exchange between a first computing device and a second computing device;
means for preprocessing the unstructured textual data and metadata at the instruction generation system by:
determining, based at least in part on the metadata, one or more portions of the unstructured textual data originating with the first computing device; and
generating, based at least in part on the one or more portions of the unstructured textual data originating with the first computing device, machine-learned model input data by removing extraneous data from the one or more portions of the unstructured textual data originating with the first computing device and tokenizing the one or more portions of the unstructured textual data originating with the first computing device;
means for executing a machine-learned model using the machine-learned model input data as input to generate operator data as output; and
means for transmitting, to the second computing device, instructions to generate an interface at the second computing device, the interface comprising an interface element comprising a graphical representation of the operator data.