SentiSquare workshop with the No-Code NLP platform

What does a “real AI” experience look like?

SentiSquare engineers/lecturers prepared 5000 texts about wine from social media and other sources as test data. Furthermore, each participant was then given individual access to the SentiSquare No-Code NLP platform so they could follow the lecturers during the workshop and test everything the platform has to offer.

Firstly, we tried the so-called “discovery mode” – unsupervised machine learning. For example, unsupervised machine learning detected the main topics in the texts where other such findings such as keywords, or so-called “clusters”, could then be displayed and visualized in the dashboards.

Next, we proceeded with the so-called “observation mode” – supervised machine learning – to train our classifier. Here we created a binary classifier to determine which texts were and were not about wine. This is because we had just used the above-mentioned discovery mode to find that topics that were not related to wine had wandered into the data, meaning these topics needed to be identified.

The trained classifier achieved a success rate of around 90%, which is an excellent result.

The detected outputs were again effectively visualized in the dashboards of our No-Code NLP platform. It was in these dashboards that participants discovered the possibilities of graphs, reports, and all output combinations. Participants could then, for example, filter the positive things being discussed on Twitter as well as the negative things mentioned by people on Facebook.

At the end of our workshop, we created a sentiment classifier that determined the sentiment across different topics about wine. Of course, we again visualized this output in our dashboards.

A well-deserved break with wine

Since our workshop – or rather the provided data – was about wine, there was only one way we could end this three-hour discussion: glasses of wine and a well-deserved break :-).

Thank you all for your participation, dedication to creating classifiers, and desire to learn how AI operates, works, and thinks :-).

We hope everyone enjoyed the workshop, and we will soon create another event for you and all other AI and innovation fans.

Thanks also to our partners IXPERTA and Katarína Kubů for co-organizing the affair.

‍