The second is the noise on the signal you want to monitor.

TG Data Set: A collection for training AI models.
Post Reply
Bappy10
Posts: 672
Joined: Sat Dec 21, 2024 5:31 am

The second is the noise on the signal you want to monitor.

Post by Bappy10 »

Social media are a curse and a blessing in this. A curse: because of the speed with which questions or rumours circulate during a crisis. As a result, information is taken out of context. A blessing: because as a government organisation you can quickly gain insight into what your environment is talking about or what it wants to know during a crisis. Of course, social media are not the only means for environmental analysis.

Targeted search or for that one grain of sand?

Four challenges
There are four challenges when we talk about environmental analysis:

The first challenge is actually: Why are you monitoring? For what purpose do you perform the monitoring? This is what we have described as the first theme in this article. Is it mainly intended for (crisis) communication or does it have a broader scope? Suppose the environmental analysis is intended to find a trigger for the occurrence of an incident (for example riots). Not an easy process. After all: where do you search? And on which words? A tool for this can be: taking the risk analysis or the risk profile of the organization as a starting point. In this way, the risks serve as a direction to set the search terms.
For example, think of fire. Brand beer. Brand manager. Brand clean. Brand is also a hamlet in the municipality of Landerd in Noord-Brabant. And De Brand is an industrial estate in Den Bosch. Fortunately, this challenge is easier. If you can find out the reason for the noise, you can also filter it out. We will give these tips later.

Thirdly, can everything be monitored? Think of the many errors in the messages or in the words. Communication instead of communication. Nuclear instead of nuclear. Only one question is relevant here: are you missing significant information? We think that this is not so bad. By the way, by going very broad, you also get more noise.

The last one is the most important: interpreting the monitoring. This is a bias that is not easy to brush away. Interpreting sarcasm or humor is already difficult in everyday conversations, let alone in online conversations about your organization.
Post Reply