On February 26, 2019, The Spindle, in collaboration with the Humanity Hub, organised Future Session #1 of the year about Data, Machine Learning and Artificial Intelligence (AI). The session was facilitated by Landscape and featured both an interesting introduction to the topic and an interactive exercise on how to use data and datasets to create AI solutions.
In a time where data might be the “oil of the 21st century,” this introductory session tried to shed light on questions such as: why is data powerful? What are the opportunities of data for organisations? How can artificial intelligence be used to create data-driven solutions?
Big data, data science and AI
The session began with a presentation by Jacob Boon, consultant at Landscape. Landscape is a company that helps organisations make use of data science in order to develop and apply data-driven and AI solutions. Jacob started by providing some definitions of the most important concepts. Big data was defined as “data that cannot be efficiently processed with traditional means.” Consequently, data science is the field that makes data fit for reading and processing by a computer, in order to be able to analyse them and eventually implement the discovered patterns to develop impactful data-driven solutions. Finally, Artificial Intelligence is a general term used to indicate the replication of intelligent human decision-making by non-humans.
Machine Learning is the next step in the process and consists of training computers/AI to use big data to learn from them and perform new tasks accordingly. The goal of machine learning is to train computers to convert an input (examples, in the form of data) into the desired output (the target). In order to develop these converting models, it is fundamental to gather data and make them speak to each other. However, linking datasets (i.e. data sources) in a working algorithm is possible only and only if one same kind of information appears in both. Accordingly, choosing the right data sources is crucial in the process.
Card game: reasoning about data
In the second half of the session, participants engaged in an interactive exercise in which they had the opportunity to put into practice what they had just learnt. The audience, divided into three subgroups, was asked to design a data-driven solution to determine how many visitors each amusement park in the Netherlands should expect on any specific day. Each group was given eight cards, each having two sides: the front featured a description of the dataset, while the back listed the exact (kinds of) data contained in the dataset. Examples of cards were: cash register sales; weather; food and drinks revenue. Participants had to deduce from the front side of the card which datasets might be useful and which not in order to provide a solution to the case. Subsequently, they could turn the cards and had to link the datasets with each other as to eventually achieve the target.
How will AI work for you?
After the interactive simulation, Jacob suggested a scheme to follow to make AI work for your organisation. First, take stock of ideas within your team on how to use data to solve some challenge that your organisation is facing. Subsequently, to choose where to start from, rank the opportunities on the basis of two criteria. Which is the extent of the expected impact? How large is the required effort? Finally, build a predictive model based on the available data and datasets, test it in a pilot and – if successful – upscale its implementation. In the end, participants were invited to share their ideas on how to make use of AI solutions for their organisation, some of which were very interesting and promising. Accordingly, a number of possible collaborations with Landscape were identified and further discussed during the closing drinks.
Here you can find the presentation that was given by Jacob Boon of Landscape.
Interested to discover even more about data, (technological) trends and their relevance for the future of development cooperation? Join us in the next Future sessions! The upcoming one will be on Earth observation & Big data analysis and will take place on Tuesday, March 26, 2019.