Python App Development Needed
Budget: $10 – $30 USD
Scenario:
DigitalPulse is a digital agency that provides social media management and analysis services for businesses in different sectors. DigitalPulse offers innovative solutions to strengthen the brand image of its clients and optimize their social media performance.
Recently, DigitalPulse has observed a significant increase in the number of social media comments from its clients. These comments contain various opinions on different topics. However, the large volume and diversity of these comments have made the analysis process complex and time-consuming. DigitalPulse aims to analyze these comments more effectively to provide more accurate and timely feedback to its clients.
Objective:
The Artificial Intelligence Team at DigitalPulse aims to develop a solution for the increasing comments. They want to group comments related to specific topics, classify them as supportive or opposing, and generate a general conclusion sentence about the topic's social media sentiment. Each output should be evaluated with appropriate metrics.
Integration:
Model and application integration is based on an event-driven architecture. Comments made in the application are transmitted as event messages and stored in a database. These comments in event messages are transferred to the model for analysis and processing. The model returns its results via event messages. During this process, the model's services can be provided via GRPC API.
Dataset:
In the topics.csv file, there are main opinion paragraphs/sentences related to different topics along with unique IDs.
In the opinions.csv file, there are opinions about different topics and their positions (claim/counter_claim/evidence, etc.). The topic_id column indicates which topic each opinion is related to.
Claim: A statement supporting a position.
Counterclaim: A statement that refutes or presents an opposing reason to another claim.
Rebuttal: A statement refuting a counterclaim.
Evidence: Ideas or examples supporting claims, counterclaims, or rebuttals.
The conclusion.csv file contains the desired conclusion sentence of DigitalPulse after analyzing the opinions on social media, restating the claims, and summarizing the opinions about the topic.
IMPORTANT NOTE: The purpose of this study is to correctly group topics with opinions and reach a conclusion. The topic_id columns provided in the dataset should only be used for validation purposes. They should not be used as a feature during testing.
A classifier and/or sentence similarity model can be used to combine topics with opinions. There is no restriction on the model. When generating conclusions, the use of a Language Model (LLM) is expected.
Input (topic.csv):
Main Opinion: On my perspective, I think that the face is a natural landform because I don't think that there is any life on Mars. In these next few paragraphs, I'll be talking about how I think that is a natural landform
Output (opinions.csv):
Relevant Opinion 1 (Claim): I think that the face is a natural landform because there is no life on Mars that we have discovered yet.
Relevant Opinion 2 (Evidence): If life was on Mars, we would know by now. The reason why I think it is a natural landform is that nobody lives on Mars to create the figure. It says in paragraph 9, "It's not easy to target Cydonia," which means it's not easy to know if it is a natural landform at this point. In all that they're saying, it's probably a natural landform.
Relevant Opinion 3 (Counterclaim): People thought that the face was formed by aliens because they thought that there was life on Mars.
Relevant Opinion 4 (Rebuttal): Though some say that life on Mars does exist, I think that there is no life on Mars.
Output (conclusion.csv):
Conclusion: Though people were not satisfied with how the landform was considered a natural landform, in all, we knew that aliens did not form the face. I would like to know how the landform was formed. We now know that life on Mars doesn't exist.
DigitalPulse is a digital agency that provides social media management and analysis services for businesses in different sectors. DigitalPulse offers innovative solutions to strengthen the brand image of its clients and optimize their social media performance.
Recently, DigitalPulse has observed a significant increase in the number of social media comments from its clients. These comments contain various opinions on different topics. However, the large volume and diversity of these comments have made the analysis process complex and time-consuming. DigitalPulse aims to analyze these comments more effectively to provide more accurate and timely feedback to its clients.
Objective:
The Artificial Intelligence Team at DigitalPulse aims to develop a solution for the increasing comments. They want to group comments related to specific topics, classify them as supportive or opposing, and generate a general conclusion sentence about the topic's social media sentiment. Each output should be evaluated with appropriate metrics.
Integration:
Model and application integration is based on an event-driven architecture. Comments made in the application are transmitted as event messages and stored in a database. These comments in event messages are transferred to the model for analysis and processing. The model returns its results via event messages. During this process, the model's services can be provided via GRPC API.
Dataset:
In the topics.csv file, there are main opinion paragraphs/sentences related to different topics along with unique IDs.
In the opinions.csv file, there are opinions about different topics and their positions (claim/counter_claim/evidence, etc.). The topic_id column indicates which topic each opinion is related to.
Claim: A statement supporting a position.
Counterclaim: A statement that refutes or presents an opposing reason to another claim.
Rebuttal: A statement refuting a counterclaim.
Evidence: Ideas or examples supporting claims, counterclaims, or rebuttals.
The conclusion.csv file contains the desired conclusion sentence of DigitalPulse after analyzing the opinions on social media, restating the claims, and summarizing the opinions about the topic.
IMPORTANT NOTE: The purpose of this study is to correctly group topics with opinions and reach a conclusion. The topic_id columns provided in the dataset should only be used for validation purposes. They should not be used as a feature during testing.
A classifier and/or sentence similarity model can be used to combine topics with opinions. There is no restriction on the model. When generating conclusions, the use of a Language Model (LLM) is expected.
Input (topic.csv):
Main Opinion: On my perspective, I think that the face is a natural landform because I don't think that there is any life on Mars. In these next few paragraphs, I'll be talking about how I think that is a natural landform
Output (opinions.csv):
Relevant Opinion 1 (Claim): I think that the face is a natural landform because there is no life on Mars that we have discovered yet.
Relevant Opinion 2 (Evidence): If life was on Mars, we would know by now. The reason why I think it is a natural landform is that nobody lives on Mars to create the figure. It says in paragraph 9, "It's not easy to target Cydonia," which means it's not easy to know if it is a natural landform at this point. In all that they're saying, it's probably a natural landform.
Relevant Opinion 3 (Counterclaim): People thought that the face was formed by aliens because they thought that there was life on Mars.
Relevant Opinion 4 (Rebuttal): Though some say that life on Mars does exist, I think that there is no life on Mars.
Output (conclusion.csv):
Conclusion: Though people were not satisfied with how the landform was considered a natural landform, in all, we knew that aliens did not form the face. I would like to know how the landform was formed. We now know that life on Mars doesn't exist.