Description
Exploring the avocado cartels topic on social media allowed us to understand how people were talking about it, focusing on the Mexican territory. The first research included the main social networks used by Mexican users to share content; we selected Facebook and YouTube as the top ones in rankings. Thanks to the nature of the two social networks, we could see how people reacted to contents through comments. We had to consider that the frequency with which people shares posts or videos is different depending on the platform and that user could not have the same reaction activity on both of them. We then searched for the same queries either on Facebook and YouTube, “aguacate carteles” and “avocado mafia,” and skimmed through the results to see the relevant ones.
The first ten results, selected, provided a good amount of useful comments for our analysis. We first noticed that there was a lot of content shared by news channels on social media, but most people did not engage in any reaction. This led us to think that people would rather share information than share their own opinion in a personal post. When it comes to commenting a post about avocado related issues, however, a good number of people take the same side and express what they think or feel towards these problems.
Therefore, the first visualization shows that many posts do not have any comment reactions and that the main discussions were nested in a small number of them. A possible insight is that people are still not so interested in talking about the issue on social networks; as we read the posts, we also understood that this probably happens because users just like and share articles they find online.
Protocol
To obtain useful results, we first set Google Incognito mode language to Spanish and country to Mexico. Then, we searched for “aguacate carteles” and “aguacate mafia” queries on Facebook and YouTube (keeping the quotation marks for the latter one for a more precise search). Since the posting frequency is higher on Facebook, we set the year 2017 for our results for this platform only. We then proceed to scrape the first ten results (if present), sorting them by relevance seeing whether they were adherent to our topic or not and if they had any comment to examine.
Listing the useful ones along with their number of comments and comment body, we could proceed to read them all and analysis their potential relevance for a mood analysis. With the information collected we designed visualization 1: in the central part, the queries we used for the search on both Facebook and YouTube are listed; a series of connections start from these points, according to the number of posts resulting from those queries.
Data
Data Source:
Google advanced RAWgraphs Statista Mexico Social Network Penetration
Timestamp: 19/10/2018
View Dataset
For this dataset, we searched for the queries used in the protocol and then we manually scraped all the information we needed; listing them in an Excel file allowed us to design the visualizations.