search

rNLKJA / Australia-Social-Media-Analytics-on-the-Cloud

2023 S1 Cluster and Cloud Computing Assignment 2
http://172.26.130.83:3000
MIT License
4 stars 0 forks source link

What scenario are we interested in #7

Open rNLKJA opened 1 year ago

rNLKJA commented 1 year ago

Think about some possible/deployable topics

RunqiuFeiKM commented 1 year ago

Topic Proposal 1: Tweet sentiment analysis across time and space How Tweet sentiment distributes geographically across Melbourne or Australia? Does the time of a day/month/year also matters? [Can we also find similar patterns from the SUDO data?]

rNLKJA commented 1 year ago

Topic Proposal 1: Tweet sentiment analysis across time and space How Tweet sentiment distributes geographically across Melbourne or Australia? Does the time of a day/month/year also matters? [Can we also find similar patterns from the SUDO data?]

When you refer the geo-space, you mean we exam the information across every suburbs or similar to what we’ve done in asmt 1?

Because based on this Ed thread #255 only 6% twitter data contains detailed geo information.

RunqiuFeiKM commented 1 year ago

Topic Proposal 1: Tweet sentiment analysis across time and space How Tweet sentiment distributes geographically across Melbourne or Australia? Does the time of a day/month/year also matters? [Can we also find similar patterns from the SUDO data?]

When you refer the geo-space, you mean we exam the information across every suburbs or similar to what we’ve done in asmt 1?

Because based on this Ed thread #255 only 6% twitter data contains detailed geo information.

Probably use the method that is most approachable or implementable for geo-space data collection. And for #255, it seems that 6% twitter data should be enough for analysis (see the final comment from Richard in the post).

RunqiuFeiKM commented 1 year ago

Many SUDO data may NOT be accessible (e.g. the psychological distress data), so we should really test the accessibility of data when browsing for it (i.e., click "Add and go to Database”).

However, Tweet sentiment analysis across time and space may still be a good research question since it can easily relate to many other topics (e.g., income, housing, environment, schooling, and recreational activities). The data for some of these topics can easily be obtained from SUDO.

@ZongchaoXie