Harnessing blended learning data to address the education and attainment gap as a result of Covid-19

2019冠状病毒病对英国长期未来最重要的影响之一是对我们年轻人的教育。据了解，封锁对学生产生了两种重大影响。首先是在失去学习的意义上的成就/进步差距。影响程度的因素可能包括特定学校的远程教育（或为弱势/关键工人的子女提供面对面教育）的质量和数量、家长在家学习的质量和数量，以及各种家庭背景因素，包括社会经济、物理环境、技术和资源的获取等。第二类影响是心理/人际/发展。年轻人，特别是那些在封锁期间没有上学的年轻人，也可能已经无法适应学校的学习环境、结构和常规；与同伴的互动；以及与老师的互动。这种失去的适应能力，加上封锁生活的其他困难，可能影响了他们恢复失去的学习能力、更广泛的社会心理发展和心理健康/福祉的能力。这些差距和因素目前是假设的，但它们在封锁期间对年轻人的影响程度尚未得到很好的理解或证明。收集这些理解和证据对于确保当前的学生群体不会成为迷惘的一代至关重要，从而对他们、更广泛的社会、国家劳动力和经济产生长期的有害影响。在这个历史时刻，我们有一个独特的机会，可以通过使用远程/混合学习数据来分析学龄儿童的学习和发展差距。在过去的几个月里，在数字学习平台上发生的教育活动从未如此之多。Sparkjar是我们世界领先的英国混合式学校学习平台，在封锁期间使用率增加了700%。在封锁期间，我们使用的两所试验田中学全程运行完整的教学时间表，交换了17万多条信息，布置了9000多个作业，学生提交了6.9万多份作业。这些数据不仅数量惊人，而且质量和完整性也是前所未有的，因为这些学校几个月来一直使用Sparkjar作为主要的交互方式。在此之前从未有过如此丰富的数据集来分析主流教育背景下的学习和行为模式。与此同时，我们正处于机器学习和大数据领域允许对大数据集进行前所未有的深度分析的阶段。我们的愿景是利用这些独特数据的力量和价值，为学校提供一个实时的仪表板，了解学生的学习、进步、行为和心理健康。我们称之为Sparkjar实时洞察（SRI）。我们将利用尖端的机器学习、专家系统、统计分析和数据可视化来发现模式和异常，并在三个粒度级别上提供实时可操作的见解：单个学生、学生群体和整个学校。与目前定期手工报告和电子表格分析的最佳实践相比，这将代表学校数据分析的巨大飞跃。

One of the most significant impacts of COVID-19 on the long-term future of the UK has been on the education of our young people.Lockdown is understood to have created two significant types of impacts on students. The first is the attainment/progress gap in the sense of lost learning. Contributing factors to the level of impact are likely to include the particular school's quality and quantity of remote education (or face-to-face provision for vulnerable/key workers' children), the quality and quantity of home learning by parents, and a variety of home background factors including socio-economic, physical environment, access to technology and resources etc.The second type of impact is psychological/interpersonal/developmental. Young people, particularly those who have not been in school throughout lockdown may have also lost acclimation to the school learning environment, structures and routines; interactions with peers; and interactions with teachers. This lost acclimatisation, alongside other difficulties of lockdown life, may have affected their ability to recover lost learning, wider psychosocial development, and mental health/wellbeing.These gaps and factors are currently hypothesised but the extent to which they have affected young people over lockdown is not yet well understood or evidenced. Gathering this understanding and evidence is crucial to ensure the current cohort of students do not become a lost generation, with long term detrimental effects on them, wider society, and the national workforce and economy.We have a unique opportunity at this moment in history to analyse learning and development gaps in school children through the use of remote/blended learning data. Never has so much educational activity taken place on digital learning platforms as over the past few months.Sparkjar -- our world-leading UK-based blended learning platform for schools -- saw +700% in usage during lockdown. Throughout lockdown, our two testbed secondary schools using have run full teaching timetables throughout, exchanged 170,000+ messages, set 9,000+ assignments, and 69,000+ pieces of work have been submitted by students.Not only is this quantity of data remarkable, but its quality and completeness are unprecedented because the schools have been using Sparkjar as their primary means of interaction for a number of months. Never before has there been a data set as rich to analyse patterns in learning and behaviour in the context of mainstream education.At the same time, we are at a point where the fields of machine learning and big data permit an unprecedented depth of analysis of large data sets.Our vision is to harness the power and value of this unique data, providing schools with a realtime dashboard of insights into student learning, progress, behaviour and mental wellbeing. We call this Sparkjar Realtime Insights (SRI). We will leverage cutting-edge machine learning, expert systems, statistical analysis and data visualisation in order to spot patterns and anomalies, and provide realtime actionable insights at three levels of granularity: individual student, student groups, and whole-school. This will represent a quantum leap in school data analysis for schools when compared to current best practice of termly manual reporting and spreadsheet analysis.