4 types of data to apply Artificial Intelligence

There are 4 types of data for Artificial Intelligence (AI). Understanding them will not only give us an idea of the data that will be necessary for our AI project, but also provides us with a very relevant vision of the use cases that we can implement.

The 4 types of data for artificial intelligence are:

1. Image data
2. Natural language data
3. Sensor data
4. Transactional data

Image data

Although humans can see a photo and immediately recognize any object, this was not easy for computers until very recently. Traditional computer programming required developers to give detailed instructions to computers on exactly what to do in any situation.

Currently, we can program computers to learn things from their own experience. This is thanks to the advances made in Machine Learning (ML), as well as the increased computing and storage capacity of computers that allows data scientists to use approaches similar to those used in the human brain (neural networks).

Some examples of use cases with image data are:

• Verification de la identidad mediante comparación facial
• Analysis del uso de equipos de protección y mascarillas en el lugar de trabajo
• Detection of thousands of objects such as brand logos

Actually, any of us are using this type of algorithms when we upload images to Google Photos or Amazon Photos. Modern cars also use this type of algorithms to detect what is happening in their environment. The most recent novelty that has attracted a lot of attention in the field of image AI is DALL-E-2, capable of creating images from text.

Natural Language Data

Natural Language Processing (NLP) is a field of artificial intelligence where computers analyze, understand, and derive meaning from human language. NLP is another example of a simple problem for humans but very difficult for traditional computing. Understanding human language is not only about understanding words, but also concepts and how they are interconnected to create meaning.

PNL is commonly used for text extraction, automatic translation, and automatic response to questions with customer service chatbots. It is also used for transcribing voice to text or incorporating realistic voices into our applications, and even for conducting "sentiment analysis", that is, understanding the mood of our customers and improving their attention.

Sensor data

Today, the proliferation of Internet of Things (IoT) both in the home and professional environments has led to almost any device we use in our homes, offices, factories, or even on our bodies being online and connected. Our cities are sensorized (smart cities) and measure traffic, air quality, noise, and all kinds of data that help improve the quality of life of citizens.

At an enterprise level, IoT has enormous implications on how we manufacture goods, provide services, sell to customers, and follow up with support. Smart factories and logistics plants are becoming increasingly automated. For example, applying artificial intelligence to sensor data enables predictive maintenance, that is, predicting where breakdowns will occur before they happen to replace and repair faulty equipment more efficiently, and even prescribing tasks to our operators to facilitate decision-making at certain points in the operational chain.

Transactional Data

Transactional data is the recorded information of the transactions carried out by our users. A transaction is a sequence of information exchange that fulfills a request, for example, a purchase in an ecommerce or a viewing on a streaming platform.

Our credit card data is transactional data. Hence, an important field for data-driven artificial intelligence is fraud detection and payment transaction analysis. By applying artificial intelligence to this type of data, we can also forecast sales in stores and predict stockouts in warehouses. Other common use cases include personalized recommendations to our users (such as the recommendations made by Amazon, Youtube, Netflix, or Spotify) as well as the implementation of personalized promotions and cross-selling.

We have seen the different types of data for AI. Starting by understanding the data is a very useful way to glimpse the type of applications that AI offers us.