The Aporta Challenge, in line with many other initiatives promoted by public administrations, could not be unaware of the great challenges we are facing in this year 2020. For this reason, its third edition, while fulfilling its usual objective of promoting the use of data and related technologies, aims to contribute to solving problems related to digital education. Without doubt, this is one of the areas in which the need to propose new innovations to ensure that the pandemic does not cause serious damage to the potential of the younger generations has been most evident.

With the slogan "The value of data in digital education", datos.gob.es is proposing an Aporta Challenge that in 2020 reward ideas and prototypes that identify new opportunities for capturing, analysing and using the intelligence of data in the development of solutions for the educational sphere at any of its stages.

Identifying a problem

If we were to approach participation in the challenge as a data science project, the first thing we would do is determine the question we would like to solve, in short, choose a problem worth working on.  In this article we propose some lines of work, but they are not restrictions, they are only intended to serve as inspiration to make it easier for us to choose an educational challenge with a great impact. We must always aspire to improve the world.

On the other hand, we can look at the large educational gaps defined by the Educa en Digital programme, which aims to complement the Digitalisation and Digital Skills Plan and to promote the digital transformation of education in Spain, making intensive use of ICT both in the classroom and in non-presential formats, and tackling specific problems thanks to developments linked to data and artificial intelligence. For each of the specific objectives we can think of a good number of issues on which we can work:

1. The provision of digital educational devices and resources. For example, how can we help ensure that access to technology is not a barrier to access to education, especially for the most vulnerable groups? how can we reduce the requirements for accessing educational programmes remotely? how can we rely on the most economical devices that are most widely available to students? etc.
2. The provision of digital educational resources, especially in relation to the previous point. On many occasions the problem we can work on does not have to be completely new, but we can find a more efficient approach to an apparently resolved issue.  For example, how can we help a teacher to better monitor a large number of students? how can we improve the security of the applications used by students through public networks? how can we guarantee the privacy of students? etc.
3. The adequacy of teachers' digital skills. In this line there are also a significant number of questions to be resolved: how can we improve the usability of tools for teachers and students? how can we promote skills related to collaboration or communication when people are not in the same physical space? how can we help STEM skills to be perceived as transversal? etc.
4. The application of artificial intelligence to personalised education, which is almost a holy grail of Education. How can we create personalised learning paths for each group of students, or better still, optimising the learning pace of each student according to their individual characteristics? how can we predict the impact of changes in programmes on the evolution of group or student learning? how can we detect and avoid gender bias in models that work on any of the above problems?

In short, with the suggestions published in the bases and a little research, it is easy to locate a good number of issues on which we can do our bit to improve digital education. Without forgetting our own experience. We have all been at least students, and perhaps also teachers, at some point.

Examining the prior art

Before we begin our work, we must consider that it is very likely that, with or without success, others have identified and proposed solutions to the problem we have chosen. From their success or failure, we can also draw lessons so reviewing the state of the art is key to focusing our project well. In relation to educational technology it is interesting to review resources such as

• The activity of educational technology start-ups in repositories such as EU-startups or the WISE accelerator.
• Awards focused on educational technology such as the prestigious Global Learning XPRIZE or the WISE Prize for Education.
• The list of more than 2.500 educational innovation projects from around the world contained in the Leapfrogging Inequality: Remaking Education to Help Young People Thrive.
• The solutions that reuse open data in the area of education and that highlight portals such as the European data portal or datos.gob.es.

As you will see, many of the projects are focused on solving problems that are mostly present in countries less developed than ours. However, the pandemic has changed the rules of the game from what we could have foreseen and is challenging us again with problems that under normal circumstances we would consider to be overcome.

Locating datasets

Open data is present in almost every problem that is solved by data related technologies and it is usually one ingredient, not the only one. The foundations of the Aporta Challenge reflect this reality and impose very few restrictions on creators, using data sources listed in datos.gob.es is not even mandatory, despite being the driving force behind the challenge. At least one set of data generated by the public administrations must be used, but it can come from any source and can play any role within the project.

To locate data related to our project we can start with the more than 1,700 datasets of the datos.gob.es data catalogue, which federates a good part of the data available in Spanish portals. In the European Data Portal we can find more than 8,000 datasets related to education from all EU countries and another 3,000 datasets from the catalogue of the European Union open data portal.

International institutions that work for the development of education such as UNICEF or the World Bank also have open data catalogues in which we can locate resources that help us in some part of our project.

The Google dataset search engine, the AWS open data registry or the Microsoft Azure datasets are resources in which we can also find datasets to enrich any data-based project.

The data catalogue of institutions such as the US Government's Institute of Education Sciences, which although focused on the United States, undoubtedly contains data of great value for measuring and understanding the impact of initiatives developed to improve education and which can enrich many projects.

Another option that we should bear in mind is that it may not be enough to solve the problem we have chosen to clean up, reconcile and transform datasets from any of the sources that are publicly and openly available. Sometimes we need to work on generating or building our own dataset. And in that case a very good option is to make it publicly and openly available so that it can be reused and improved by others.

Defining the product

Finally, we have to think about the best way to deliver the result of our work so that it can be used by its recipients and have the impact we want. The options are multiple and again the bases do not impose restrictions. Some possibilities could be:

• Mobile Apps: The enormous penetration of the iOS and Android platforms means that any product we build for these platforms and publish in their respective stores is guaranteed to have a huge potential diffusion. In addition, there are options to carry out multiplatform developments and even to carry out developments with little (low-code) or no (non-code) software development knowledge.
• Websites: Web applications are probably still the most common mechanism for making a project of any kind available to society in general. The advances in managed services of the large cloud providers and the facilities they offer to make infrastructure available for free mean that it has never been easier to start a project. It is also possible to use non-code platforms such as appypie or low-code platforms such as Appian to reduce the initial barrier if we do not have a software developer on the team.
• Artificial Intelligence Algorithms: It is increasingly common for a data-based project to be delivered in the form of an automatic learning model or artificial intelligence. For example, Amazon AWS offers the possibility to list algorithms like Microsoft Azure in its Machine Learning Marketplace so that they can be consumed by other applications.
•  Stories and Visualizations: Sometimes the best way to deliver results is through a visualization or a DataStory that allows you to communicate the result of your work. For this purpose, there are multiple options that range from the utilities that incorporate most of the generic Business Intelligence tools such as Tableau to others specialised in spatial location such as the Spanish Carto.

We wish all participants good luck and encourage you to work on a challenge that has a great impact on society.

Content prepared by Jose Luis Marín, Senior Consultant in Data, Strategy, Innovation & Digitalization.

The contents and points of view reflected in this publication are the sole responsibility of its author.