Contents of this page:

Choosing a social media platform

Once you have a problem statement, you need to decide what data from social media is going to help achieve your objective and answer your research questions.

To identify relevant data, you need to first decide what social media platforms are relevant to your problem statement. Ask these questions:

• What is the penetration of each social media platform? You can usually find this out with a little desk research.
• What are different social media platforms used for? You can find some of this from desk research, but talking to people from the context is likely to get you further. This is important if some platforms are e.g. more political, or used differently depending on gender or age, for example.

Phoenix currently supports gathering data from the following social media platforms:

There are some limitations to what data Phoenix can gather, which are detailed in the next section.

Deciding on search parameters

Once you have decided on the social media platforms you will focus on, you will need to brainstorm on some search parameters for each social media platform. You will iterate and experiment on the Phoenix platform to refine these search parameters, but it is good to have a brainstormed a starting point. Search parameters include:

• What length of time you want to gather posts / comments for.
• What list of accounts you want to gather posts / comments from. Jump to: tips for brainstorming a list of accounts.
• What list of keywords help narrow down to the posts / comments you are most interested in.

<aside> 💡 Once you have an idea for your starting point with search parameters, you may want to try them out using the gather options on the Phoenix platform, and then iterate.

</aside>

Tips for brainstorming a list of accounts

1. Start by brainstorming who are people or groups that have a public presence on the platform(s) of choice, are relevant to the context and are talking about what you are interested in (based on the problem statement). You might want to think of public officials, groups of citizens, organizations, influencers, and more.
2. Use the search function on the platform to find these people or groups; then decide whether their account should be added to your data sources by considering:
   1. Whether they are from the geographic area you are interested in. Geography can be in their account name, or there may be other location data (e.g. you can filter Facebook pages by the location of their admin)
   2. How many followers they have. Accounts that have over 1000 followers are more likely to have a trending conversation as the followers engage with the content. If you however find an account with very relevant posts but with less than 1000 followers you may still want to include it.
   3. How often they post. Accounts that only post every few months are unlikely to be useful. Accounts that post 20+ times per day are likely to overwhelm your data (this is common for media pages).
   4. What they post about. Remember that you are looking for accounts that post quite often about content relevant to your problem statement.
   5. How people react to their posts. Look at whether people are commenting on, liking or sharing these posts. You ideally want accounts that have a decent level of engagement.