Making new tools part of the plan: Meet the winners from Hacking EdPlanning

01 February 2021

hackathon.png

Conan Project
The Conan project working through the night of 30-31 January 2021 for Hacking EdPlanning.

Software engineers, data scientists, and programme designers joined forces with educational planners this weekend for the first IIEP-UNESCO Hackathon -- Hacking EdPlanning -- organized in collaboration with the tech for good network organization Latitudes. Bringing together 112 people from 10 countries, this energy-filled event confronted six challenges in educational planning and put forth new digital prototypes – in just 48 hours – to help countries plan for a brighter educational future. 

“We believe that innovation, technologies and techniques can substantially contribute to achieving equitable, quality education systems by 2030,” said IIEP’s Director Suzanne Grant Lewis at the close of Hacking EdPlanning on 31 January 2021.

 

At the end of the weekend, a jury awarded two prizes: one for impact, and one for creativity and innovation. A third prize - the Audience Award - was selected by the public.

Winner for impact

IIEP-UNESCO Hackathon Winner for Impact

EdVis: How might we help education planners more easily identify and intervene in regions where learners are performing poorly, using learning assessment data?

This was the challenge for team “Dehta”, made up of two data scientists, three developers, and six designers based in Canada and a project manager in India. This team - winner for Impact - consisted of 12 members who all recently graduated from the tech bootcamp at BrainStation. “This was a really good opportunity to design a solution for a real world problem,” said product designer Henry Mai. Fellow team member Aletheia Délivré, an Education Technologist at AlphaPlus in Canada, agreed: “We got to tackle a problem that pertains to sustainable development. Normally, this doesn’t get to benefit from so much brain juice in the tech space, so this was a challenge but also hugely rewarding.”

With a user-friendly, interactive interface, EdVis makes learning assessment data easy to understand for education actors so that they can target policy action. Users can visualize learning outcomes by province and link baseline data with factors that influence learning outcomes. This data-driven team had impact on their radar from the start of the hackathon: they needed to first interpret a colossal amount of data and figure out to make a well organized and intuitive interface. Nicholas Fan, a product designer and team member, said the project successfully lays the groundwork for further development: “We did this in just a weekend, imagine how much a team could do in a few months.”

Watch a short video:

Tied Innovation and Creativity Award

Conan Project: a tool for exploring school inspection data

How can a ministry of education get an overview of the topics covered in school inspection reports? As these documents are written in the thousands and generally in free text form, it is difficult to analyze the results on a large scale. This is precisely the challenge that the Conan Project team sought to solve. "We wanted to represent the data in a useful way; to put the technology to work for the end users," the team explains. Made up of a web developer, a data architect, a computer engineer and several data scientists, including a student, the seven members based in France combined their expertise for the fun of it.

"Everyone had an indispensable role to play in making the tool work, even those who only stayed for a few hours: it was quite magical. We didn't all know each other and this hackathon is a first for most of us. It's a great experience," says the team that won the Innovation and Creativity Award.

Watch a short video:

On the basis of a keyword entered in the interface - such as 'absenteeism' or 'school material' - the user instantly visualizes the position of the schools linked to the most relevant reports on this theme. He can thus identify any territorial disparities. The display on the map also allows the user to "take the temperature" of the inspection reports on the topic in question: green when the content is judged positive, red when it is negative. 

Winner of the Innovation and Creativity Prize and the Audience Award

EduChex: Catching ghost teachers through Natural Language Processing (NLP)

80,000 ghost teachers in India. 100,000 in Pakistan. 2,500 in Niger and Sierra Leone. Thousands more undetected. To confront this reality common to countries worldwide, the team “Timezones” created a solution using natural language processing to help ministries identify teachers who are no longer in classrooms, yet remaining on payroll. “Our solution uses natural language processing to streamline and quicken teacher data checks for ministries,” said Amreen Poonawala, from the Harvard Graduate School of Education, in her team’s winning pitch. “It is user-friendly. It is a plug-in that integrates into the ministries existing workflows and automates the matching of teacher entries.”

Watch a short video: 

“It was nice to work with so many talented folks with so many deep functions and contribute to something that would have real world application,” said Pierce Henderson from the Harvard Graduate School of Education. Along with five other members - spanning three time zones and hailing from Canada, Germany, and the United States -  this team had a mix of backgrounds in education, business and technology. “We were essentially bringing together two teams. We had a back-end team that works really well together, as well as a front-end one, and I think it is that mesh,” said Ashley Etemadi, also from Harvard. “The team is always the most important thing.” 

 

Tied Public Prize

Athena: when artificial intelligence illuminates inspection reports

Their names are Coline, Luca, Alexandre, Paul, Valentin, Tom, Roman, and Kilian and they are all students at Epitech school in France. The PoC team spent the weekend designing the Athena digital solution. Their objective: to automatically analyze the information contained in nearly 4,000 inspection reports from Irish schools. Based on the files made available to them, the team was able to render the "educational landscapes" of Ireland on a map.

Among its assets, the Athena prototype is based on a particularly innovative deep learning model called: ”The Transformer". "We didn't expect to work on such advanced technologies. This is the first time we have used such an artificial intelligence model. This proves the complexity of the needs of the education sector," explains Tom Chauveau. Engineer and doctoral student Nefeli Paparisteidi, a member of Epitech's pedagogical team, provided valuable support to the team. In particular, she drew up recommendations for Ministries of Education seeking to use Natural Language Processing (or NLP) technology to exploit textual data from inspection reports.

Watch a short video:

    

This prototype has numerous development prospects. According to the team, the proposed solution could be adapted to the contexts of other countries, in particular by integrating other languages.

"We worked day and night on this project. The IIEP organizers encouraged us, even at four in the morning. We are all between 19 and 20 years old,” said the PoC team. “There is nothing more motivating than working on useful projects that can improve the quality of education around the world.”.

Learn more about the Athena prototype.