Xiaolei Zhang Jul 19, 2014 11:22
Member
|
Thanks for the proposal! It will be a great job that collect huge data of people's perception about climate change, which will be very helpful for people to make decision from government, industry, and academia. At the same time, as climate change is still a topic under debate, the main difficulty probably is the variable data sources, and how to distinguish them. It will be important to distinguish comments into different levels, for example, if 8 people say something, and only 2 people have opposite ideas, we cannot conclude what the 8 people say is the truth, many times, the truth is with the other 2 people. And also sometimes, perceptions from some special social section e.g. government play an more important role to lead the main trend of people's idea. Looking forward to see this idea can be realised!
|
Nafisa Binti Jul 20, 2014 12:15
Member
|
Why do you use big data????
|
Ashraful Islam Jul 20, 2014 02:52
Member
|
The three phases on workflow in this project are properly chosen. Online social media will surely be better real time interactions in collecting public opinion in larger volumes and will be very effective in phase-I of this project. Moreover, strategies set forth in phases II and III are well thought and will be fruitful. Good luck for the team.
|
Fahim Hassan Jul 20, 2014 02:26
Member
| Proposal contributor
@zhangxiaolei0208, Thanks for your feedback. For our analysis, we are taking the "difference of opinion" into account. The nature of the analysis is not going to draw any conclusion from the web narratives, instead it will demonstrate the ways people express their opinion on web.
Instead of general frequency counts, inferential statistics, which helps to derive information about the population from the sample data, will be applied to analyze the data.
|
Fahim Hassan Jul 20, 2014 11:56
Member
| Proposal contributor
@bintinafisa, thanks for your question. There are multiple reasons for using the "Big Data" technique. I am highlighting the key reasons:
For forecasting models, higher the volume of data, better the analysis. By collecting huge volume of data, we can predict with greater accuracy. Traditional ways (SQL, relational database) are not very efficient and trustworthy to handle such high volume of data.
|
Alam Hossain Jul 31, 2014 07:23
Member
|
You have selected an excellent methodology to represent the public opinions on climate change. Good luck for the TEAM!!!
|
Yang Xiao Aug 1, 2014 12:25
Member
|
Following the question by zhangxiaolei0208 and the answer to it, I would ask: if "The nature of the analysis is not going to draw any conclusion from the web narratives, instead it will demonstrate the ways people express their opinion on web.", then what kind of content you will create "that are persuasive and effective in making lasting impressions on people’s mind" and how do you make sure these contents are reflecting the reality and the truth?
|
Fahim Hassan Aug 5, 2014 01:26
Member
| Proposal contributor
Thanks alam90119 for your inspiring words!
|
Fahim Hassan Aug 5, 2014 03:18
Member
| Proposal contributor
@ skykeepon, Thank you for your question.
To create a lasting impression on consumer psyche, we need to understand their behavior. I will illustrate a hypothetical example here to clarify.
***Example- sustainable coffee farming***
Imagine a promotional video of a coffee farm where the manufacturer emphasizes on the eco-friendly production process. The global coffee market is valuing it by defining industry standard to get verifications (i.e. rainforest alliance certificates, labels). These certified coffee's usually come with a high price tag.
Now the question is - how do you analyze the consumer's willingness to pay for the mark-up?
In earlier days, marketing department conducted consumer surveys and matched it with the price-elasticity/demand. However, surveys can have response bias as well as other limitations (framing of questionnaire, sampling bias etc.)
To overcome these issues, we are proposing to analyze the behavior that is revealed on their opinions, like the comments they made on that video. The comments can reflect the frustration of the price point, availability of the eco-friendly product etc.
Now imagine extracting a huge volume of these opinions/comments from a number of videos on Youtube. This huge dataset can reveal the way consumers value coffee from sustainable farms. Mining the text can reveal their true preferences as well as how they interpret the labels/certificates of the products.
This information is valuable, and not only for a particular company but for the coffee industry as a whole (which is over 80 billion dollar industry according to Forbes, May, 2014). The industry can have a better understanding of the growing concerns related to deforestation and sustainable change and act accordingly.
At the same time, we can understand the level of interaction with the help of memes, viral videos and other internet culture. With growing trend of online shopping, these can be used as tools to raise awareness among people about fair-trade, sustainable farming and quality coffee.
Once again, we are focusing on quantifying the expressed opinions, instead of asking customers to rank order their preferences. It's the approach that makes our proposal unique.
----
You can substitute "coffee" with "tea" or "palm oil" and the context will remain the same.
Hopefully, it answers your question. Let me know if you have any other queries or feedback.
Thanks
|