Granted Patent B2

US 12,572,406 · App. 17/956,719 · Granted Mar 10, 2026

Inventors: Ankesh Khemani (Bangalore, IN); Akshar Prasad (Bangalore, IN)

Assignees: ATLASSIAN PTY LTD.; ATLASSIAN US, INC.

G06F11/0784G06F11/0706G06F11/0721G06F11/0751G06F11/076

Abstract

Claims (95)

1 . A computer-implemented method, comprising:

collecting, from a server of a social media platform, a plurality of content items meeting a first set of relevance criteria, wherein:

the first set of relevance criteria comprises:

a first criterion that is met when each content item of the plurality of content items was at least one of created and updated at the social media platform during a first time period;

a second criterion that is met when each content item of the plurality of content items includes respective textual data including a reference to a software application; and

each content item of the plurality of content items includes respective interaction information; and

for each content item of the plurality of content items:

analyzing the respective interaction information to determine a respective weight factor;

processing the respective textual data, using a sentiment analysis engine, to obtain a respective sentiment score, the sentiment analysis engine including a trained machine learning model;

applying the respective weight factor to the respective sentiment score to obtain a respective weighted sentiment score; and

generating a geo-spatial cluster map comprising a plurality of geo-spatial clusters corresponding to a plurality of locations associated with users of the software application; and

determining a first geo-spatial cluster of the plurality of geo-spatial clusters that is associated with a subset of the plurality of content items, including:

for each content item of the subset of the plurality of content items:

in accordance with a determination that a respective content item of the plurality of content items includes geolocation data, determining, based on the geo-spatial cluster map, a geo-spatial cluster of the plurality of geo-spatial clusters closest to the geolocation data for the respective content item of the plurality of content items, the determined geo-spatial cluster corresponding to the first geo-spatial cluster;

updating a content item descriptor associated with the respective content item of the plurality of content items with an indication of the first geo-spatial cluster; and

aggregating the weighted sentiment scores of each content item of the subset of the plurality of content items to obtain a combined sentiment score for the first geo-spatial cluster;

in response to the combined sentiment score for the first geo-spatial cluster not exceeding a sentiment threshold, determining that a localized incident associated with the first geo-spatial cluster has not occurred; and

collecting, from the server, a first content item meeting a second set of relevance criteria and including first interaction information, wherein the second set of relevance criteria comprises:

a third criterion that is met when the first content item was at least one of created and updated at the social media platform during a second time period;

a fourth criterion that is met when the first content item includes first textual data and a reference to the software application; and

analyzing the first interaction information to determine a first weight factor;

processing the first textual data, using the sentiment analysis engine, to obtain a first sentiment score;

applying the first weight factor to the first sentiment score to obtain a first weighted sentiment score;

determining that the first content item is associated with the first geo-spatial cluster;

updating a first content item descriptor associated with the first content item with an indication of the first geo-spatial cluster;

aggregating the first weighted sentiment score and the combined sentiment score to obtain an updated combined sentiment score for the first geo-spatial cluster; and

in accordance with a determination that the updated combined sentiment score for the first geo-spatial cluster exceeds the sentiment threshold:

identifying a localized incident occurrence corresponding to the first geo-spatial cluster and relating to the software application; and

generating an alert message for a user, the alert message including:

an indication of the localized incident occurrence;

a geographical region associated with the first geo-spatial cluster;

at least one link to at least one content item associated with the first geo-spatial cluster; and

causing to be displayed, in an interface of a monitoring system associated with the user, the alert message.

2 . The computer-implemented method of claim 1 , wherein each sentiment score is associated with a respective sentiment label indicating a positive sentiment or a negative sentiment.

3 . The computer-implemented method of claim 2 , wherein:

determining the localized incident occurrence includes determining whether at least a first threshold number of the plurality of content items were created or updated within a threshold timespan; and

each of the at least the first threshold number of the plurality of content items are associated with sentiment scores exceeding a threshold score.

4 . The computer-implemented method of claim 2 , wherein:

determining the localized incident occurrence includes determining whether at least a second threshold number of the plurality of content items were created or updated within the first time period; and

each of the at least the second threshold number of the plurality of content items are associated with sentiment scores exceeding a threshold score.

5 . The computer-implemented method of claim 4 , wherein at least one of the second threshold number of the plurality of content items and the threshold score is a dynamic threshold value that varies over time.

6 . The computer-implemented method of claim 2 , wherein analyzing the interaction information includes determining a frequency of interactions with the content item.

7 . The computer-implemented method of claim 1 , wherein determining the localized incident occurrence further comprises determining whether a third threshold number of content items of the subset of the plurality of content items are assigned to a common geo-cluster.

8 . The computer-implemented method of claim 1 , further comprising:

segmenting a content item of the plurality of content items into two or more segments if the textual data of the content item includes more than a threshold number of sentences;

receiving, from the sentiment analysis engine, a segment sentiment score for each segment of the two or more segments; and

aggregating the segment sentiment scores for each segment of the two or more segments to determine the sentiment score for the content item.

9 . A non-transitory computer readable medium comprising instructions, which when executed by a processing unit, cause the processing unit to:

collect, from a server of a social media platform, a plurality of content items meeting a first set of relevance criteria, wherein:

the first set of relevance criteria comprises:

a first criterion that is met when each content item of the plurality of content items was at least one of created and updated at the social media platform during a first time period; and

a second criterion that is met when each content item of the plurality of content items includes respective textual data including a reference to a software application; and

each content item of the plurality of content items includes respective interaction information; and

for each content item of the plurality of content items:

analyze the respective interaction information to determine a respective weight factor;

process the respective textual data, using a sentiment analysis engine, to obtain a respective sentiment score, the sentiment analysis engine including a trained machine learning model;

apply the respective weight factor to the respective sentiment score to obtain a respective weighted sentiment score;

generate a geo-spatial cluster map associated with the software application and including a plurality of geo-spatial clusters corresponding to a plurality of locations associated with users of the software application; and

determine a first geo-spatial cluster of the plurality of geo-spatial clusters that is associated with a subset of the plurality of content items, including:

for each content item of the plurality of content items:

in accordance with a determination that a respective content item of the plurality of content items includes geolocation data, determine, based on the geo-spatial cluster map, a geo-spatial cluster of the plurality of geo-spatial clusters closest to the geolocation data for the respective content item of the plurality of content items, the determined geo-spatial cluster corresponding to the first geo-spatial cluster;

update a content item descriptor associated with the respective content item of the plurality of content items an indication of with the first geo-spatial cluster; and

aggregate the weighted sentiment scores of each content item of the subset of the plurality of content items to obtain a combined sentiment score for the first geo-spatial cluster;

in response to the combined sentiment score for the first geo-spatial cluster not exceeding a sentiment threshold, determine that a localized incident associated with the first geo-spatial cluster has not occurred; and

collect, from the server, a first content item meeting a second set of relevance criteria and including first interaction information, wherein the second set of relevance criteria comprises:

a third criterion that is met when the first content item was at least one of created and updated at the social media platform during a second time period;

a fourth criterion that is met when the first content item includes first textual data and a reference to the software application; and

analyze the first interaction information to determine a first weight factor;

process the first textual data, using the sentiment analysis engine, to obtain a first sentiment score;

apply the first weight factor to the first sentiment score to obtain a first weighted sentiment score:

determine that the first content item is associated with the first geo-spatial cluster;

update a first content item descriptor associated with the first content item with an indication of the first geo-spatial cluster;

aggregate the first weighted sentiment score and the combined sentiment score to obtain an updated combined sentiment score for the first geo-spatial cluster; and

in accordance with a determination that the updated combined sentiment score for the first geo-spatial cluster exceeds the sentiment threshold:

identify a localized incident occurrence corresponding to the first geo-spatial cluster and relating to the software application; and

generate an alert message for a user, the alert message including:

an indication of the localized incident occurrence;

a geographical region associated with the first geo-spatial cluster;

at least one link to at least one content item associated with the first geo-spatial cluster; and

cause to be displayed, in an interface of a monitoring system associated with the user, the alert message.

10 . The non-transitory computer readable medium of claim 9 , wherein:

each sentiment score is associated with a sentiment label indicating a positive sentiment or a negative sentiment by the sentiment analysis engine.

11 . The non-transitory computer readable medium of claim 10 , wherein:

determining the localized incident occurrence includes determining whether at least a first threshold number of the plurality of content items were created or updated within a threshold timespan; and

each of the at least the first threshold number of the plurality of content items are associated with sentiment scores exceeding a threshold score.

12 . The non-transitory computer readable medium of claim 10 , wherein:

determining the localized incident occurrence includes determining whether at least a second threshold number of the plurality of content items were created or updated within the first time period; and

each of the at least the second threshold number of the plurality of content items are associated with sentiment scores exceeding a threshold score.

13 . The non-transitory computer readable medium of claim 12 , wherein at least one of the second threshold number of the plurality of content items and the threshold score is a dynamic threshold value that varies over time.

14 . The non-transitory computer readable medium of claim 10 , wherein the weight factor is based on a frequency of interactions associated with the content item.

15 . The non-transitory computer readable medium of claim 9 , further comprising instructions which when executed by the processing unit cause the processing unit to determine the localized incident occurrence if a third threshold number of content items of the subset of the plurality of content items are assigned to a common geo-cluster.

16 . The non-transitory computer readable medium of claim 9 , further comprising instructions which when executed by the processing unit cause the processing unit to:

segment the textual data of each content item into two or more segments if the textual data of the content item includes more than a threshold number of sentences;

process, using the sentiment analysis engine, the two or more segments individually to obtain a segment sentiment score for each segment of the two or more segments; and

aggregate the segment sentiment scores for each segment of the two or more segments to determine the sentiment score of the content item.

Assignments (1)

Continuity (1)

References Cited (13)