Sometimes we need to reach people for marketing, advertising, rallying their support for a cause or whatever reason. Most of the time we have no idea on what somebody thinks or feels unless they write about it. But what if by looking only at a general sample of their writings, unrelated to the topic us as marketers want to sell, we can detect the opinion or sentiment of each individual person.
In the current world, most people write about whatever happens in their everyday life via social networks: Facebook, Twitter, LinkedIn, etc. Of course the topics they talk about are very specific, and probably not relevant for a marketer to target them. This is where NLP (Natural Language Processing) and Machine Learning techniques come into consideration.
Imagine a scenario where we are trying to sell apples to people, but we don’t want to just go door by door offering them. We would like to know who is more likely to buy our apples and then go directly to these highly profitable targets, while minimizing the time and money we need to reach them. Now, these people may be talking in Facebook about their lives, their dogs, family and friends. Using NLP we can figure out that people who talk a lot about dogs, actually like apples a lot too, even when they never mention them. Thus we now know to target known dog lovers, and our apples will be sold.
This graph shows how topics written in Cyrillic script are related to each other, and how they actually connect people. Allowing us to target possible customers by jumping various topics from dogs to apples.