Understanding your political social media ecosystem is essential before beginning your analysis. What are the critical questions that you should ask?
Your project will include several phases:
Here are some sample timelines based on real projects to help you get started with project planning:
See below in section Section 6, “What Timeframe?”, for more details on how to select your overall timeframe.
Conducting a risk assessment prior to starting your work may help your team understand the current situation and narrow your scope. DRI has developed a Risk Assessment framework. See our User Guide to get started:
When getting started, there are many options regarding what can be monitored, but it is not possible to monitor it all. With an ocean of opportunities, you must somehow limit these choices into a smaller selection.
To help you decide, you must keep in mind possible restraints while actively choosing your priorities. For example, if your organizational mandate is online campaign finance transparency, you might want to focus on monitoring political advertisements. You might want to monitor the top three social media platforms in your country but will have to look only at the top posts or add more analysts to your team.
Once you have selected your platform of interest, there are two main starting points when working with social media data: (a) accounts and (b) keywords. This means, at the very beginning you will need to create lists of relevant accounts to monitor and/or specific topics and keywords to search for.
Using the actor-based approach during an election, you could create a list of the relevant accounts for politicians, political parties, media outlets and political influencers. With the keyword-based approach, you could put together a list of relevant topics with relevant keywords. For example, if you identify “environment” as a topic, you would want to search posts for keywords like “#fridaysforfuture” or “extinctionrebellion”.
You can also combine these two approaches by including all posts from specific actors in addition to posts discussing certain topics. This might be particularly appropriate during elections. For example, you could monitor all known candidates, political parties and media outlets but also include posts discussing the election (e.g. #Election2020) and critical topics.
By now you should be able to answer these five key questions before beginning your monitoring.
Click on each question to check if you are ready:
Some phenomena may be particularly important based on your organizational mandate or a risk assessment of your country context:
Coming up with a list of topics is important so you can zero in on the most important political discussions in your country.
How can you come up with this list of topics and keywords?
Based on a list of topics, you can come up with a list of keywords to filter down your selection of social media data. When moving from topics to keywords, this process is experimental. You will need to see if some keywords are too general and returning unrelated social media posts.
Identify key actors by:
As a starting point, you can consider the “official” political actors in your given context:
Considering “unofficial” political actors is also important, especially when monitoring for harmful phenomena:
If you are analyzing an election, the timeframe will depend on your electoral period. For example, some countries begin campaigning over a year before an election. On the other hand, if a snap election is called then monitoring will need to begin as soon as campaigning begins. It may be important to consider:
Furthermore, DRI recommends looking at online democratic discourse beyond the electoral period. Why Monitor Social Media shows that manipulation happens around the clock, particularly:
Choose your most relevant platform by considering the following:
Upon choosing your most relevant platform, data access and availability may limit your platform selection process: