It is known that Facebook in its beginnings was intended exclusively for university students at Harvard who could communicate with each other and exchange information. Later, Facebook expanded to other physical and legal persons and has long ceased to be just a social network for the exchange of private messages and posts. The vast majority of companies today use this social network, among others, to support the business.
Businesses usually have a Facebook page on which they publish news about services, promotional activities and enable communication with end users. So why not use this information to get additional insights into the behavior of your users and reactions to services and/or promotions.
What data can be collected?
With the Informatica Big Data Management tool that has a built-in powerful Exchange Connector for Facebook, it is possible to retrieve all posted posts, whether it’s a video, a picture, or an ordinary status, information about which users commented on the same posts, and what time, content comments (text and emoticons), leaks, tags and many other useful information. The data can then be retrieved into a relational data model over which various metrics can be drawn and various analyses made.
Examples of metrics that can be obtained from the retrieved data:
- Most Popular Posts
- Number of likes per month/post/ comment
- Number of comments per day / month / post / comment
- Number of shares per day/month/ post
- Engagement (likes, comments, share)
With the Tableau data visualization tool, you can create intuitive and interactive dashboards with Facebook Analytics. In Figure 1, there is an example of reports on the most commercially available posts, depending on the needs of choosing the first 5, 10 or N posts.
We can observe and engagement depending on the day of the week so that they can see when users are most active and decide when to publish something. The example shown is shown in Figure 1.
Facebook has developed a set of reports in which owners of Facebook pages have insights, within these reports, among other things, the above metrics. These reports provide a fairly good insight into the engagement of users, but these are predefined reports that cannot be adapted to the needs and requirements.
Applying Facebook data
If all data from a Facebook page is retrieved and converted into a relational data model, it can be used for:
Linking data to existing Datawarehouse
I see the greatest potential in linking Facebook data and other data from sources or data warehouse. For example, it is possible to link DWH Customers with Facebook users (I refer to users who have ever liked, commented, or tagged within a post/comment on that Facebook page). In this way, insights from social networks can be seen in the context of other data.
For example, if we connect a custom DWH and Facebook page user, they can easily get information about their experience with the service, contentment or dissatisfaction or potential departure that has been pointed out on the Facebook page by the service provider. In this way, data from DWH can be further enriched by social networking data. The combination of these data offers the possibility to get interesting insights into user behavior.
One also very important analytic that can be done on the collected Facebook data is the sentiment analysis of comments. This approach can assess whether a user’s comment is positive, negative, or neutral, which can be done using the dictionary. What needs to be done first is to normalize the text, or break the sentences from comments to words and phrases. The point is that these words and phrases are valued and in the end, the sentiment score of each comment is calculated. Using the sentiment analysis, it is possible to obtain a relevant reaction, for example, the launch of a new product.
Create a network of users
The term SNA (Social Network Analysis) is all less familiar to all, and it is actually an analysis of social relationships and relationships. From the received data from Facebook, user connections, events, and many other relationships can be displayed. SNA helps in the discovery of hidden links and influence among entities. Often, with the help of SNA, potential churners are discovered, and applications can be used for various marketing purposes.
Ad hoc analysis
If there are prepared and structured Facebook data in the tables at the base, the company always has the ability to ask for any ad hoc analysis depending on the needs at that moment.
Searching for topics on the dashboard
As noted above, the analyst will always be able to make an analysis with the SQL query using the SQL query, but if we want a business user with an insight into the dashboards only has the option of creating flexible reports depending on which topic is currently relevant, Tableau has an excellent the ability to create a Search button on the dashboards in which a specific term is entered according to the request, and the related data is displayed on the dashboard.
The potential of using data from Facebook
For a service provider, such as a bank or a telecom, it is all about attracting new customers and keeping existing ones. It is therefore important that existing users have the experience and service they want. In order to do this, it is necessary to understand what exactly the users want and that there is an exceptional potential for diverse data from social networks.
If properly used, Facebook data can serve for the purpose of:
- improvement of user satisfaction
- improvement of services
- churn prediction
- customer acquisition
- promotion improvements
Analytics obtained from Facebook data can be used to define further strategies for achieving business goals.