Twitter topic modeling
Budget: $30 – $250 USD
Hello,
For my research project, I have downloaded tweets from a set of CEOs using Twitter API. I would like to carry out topic modeling on these tweets. In particular, I would like to categorize tweets into following five topics.
1) Financial tweets: These tweets contain information about the companies' performance. Some of the keywords would be earnings, profits, revenue, sales growth, etc.
2) Non-financial business tweets: These tweets contain information about the companies' business that are not directly mentioning performance. For example, introduction of a new product, updates on meetings with customers and/or suppliers, day-to-day operations, and etc. The idea is that these tweets would provide information to the investors even though they do not directly talk about performance.
3) ESG or CSR related tweets: These tweets contain executives' discussion related to ESG/CSR related topics. For example, tweets about climate change, work place culture, and. etc.
4) Political tweets: These tweets contain executives' discussion related to political events. For examples, tweets related to black lives matter movement, anti-vaccine, LGBTQ rights, and etc.
5) Other tweets: These tweets contain tweets that are not categorized into above four categories. These tweets would mainly be related to their personal lives that would not be informative to investors.
Key goal of the project!
For my project, reducing noise in topic categorization regarding non-financial business tweets is crucial. Based on the industry that executives operate, certain topic can be considered non-financial business tweets or other tweets. For example, if Pfizer CEO tweets out about COVID-19 vaccine trial success, it should be considered non-financial business tweet. However, if Tesla CEO tweets out about the same information, it should be classified as other tweets.
Basically, I would need a model that can reflect the type of business that the CEO operates in to reduce noise in categorizing non-financial business tweets.
I would be happy to discuss the project in more detail.
I am open for a consultation to gauge the size of the project.
Many thanks in advance!
For my research project, I have downloaded tweets from a set of CEOs using Twitter API. I would like to carry out topic modeling on these tweets. In particular, I would like to categorize tweets into following five topics.
1) Financial tweets: These tweets contain information about the companies' performance. Some of the keywords would be earnings, profits, revenue, sales growth, etc.
2) Non-financial business tweets: These tweets contain information about the companies' business that are not directly mentioning performance. For example, introduction of a new product, updates on meetings with customers and/or suppliers, day-to-day operations, and etc. The idea is that these tweets would provide information to the investors even though they do not directly talk about performance.
3) ESG or CSR related tweets: These tweets contain executives' discussion related to ESG/CSR related topics. For example, tweets about climate change, work place culture, and. etc.
4) Political tweets: These tweets contain executives' discussion related to political events. For examples, tweets related to black lives matter movement, anti-vaccine, LGBTQ rights, and etc.
5) Other tweets: These tweets contain tweets that are not categorized into above four categories. These tweets would mainly be related to their personal lives that would not be informative to investors.
Key goal of the project!
For my project, reducing noise in topic categorization regarding non-financial business tweets is crucial. Based on the industry that executives operate, certain topic can be considered non-financial business tweets or other tweets. For example, if Pfizer CEO tweets out about COVID-19 vaccine trial success, it should be considered non-financial business tweet. However, if Tesla CEO tweets out about the same information, it should be classified as other tweets.
Basically, I would need a model that can reflect the type of business that the CEO operates in to reduce noise in categorizing non-financial business tweets.
I would be happy to discuss the project in more detail.
I am open for a consultation to gauge the size of the project.
Many thanks in advance!