Members of our team recently spent a weekend in London with an amazing charity called DataKind UK. Along with a team of volunteer data experts hosted by DataKind, we set out to see what insights we could find from the data we have available on our platform. Here, we share our experience of working with DataKind, and what insights we were able to gain from the team of data scientists.
Our partnership with DataKind UK
DataKind is a UK charity that aims to transform the impact of charities through data science so that our work can have more of an impact. They connect charities with data scientists who volunteer their time and expertise with us for free.
As a tech-based charity, our platform is rich with data. This gives us crucial insights into how our programme is used and how it can be improved. In particular, we are keen to have the insights to advise teachers using our platform on how they can support disadvantaged families to engage in their children’s education – which is at the heart of our mission. We also aim to share our insights with the wider education sector.
Thanks to our tech team at Poteris, all of the data provided to DataKind had been completely anonymised and stripped back to just those essential fields needed. This data was then given to our team of data scientists to analyse.
What did we get up to over the weekend?
Day 1 was all about getting to know the other charities there at the event, meeting the data experts who chose to work with us, and the delicious free food.
We were joined by ParentKind, a charity which helps give parents a voice in their children’s education through their network of Parent Teacher Associations, and Youth Sport Trust who aim to improve young people’s education and development through sport and play.
Once the data experts had chosen which charity they would like to work with – the data diving began! Our team of data experts got busy building programmes to analyse our data.
By day 2 some incredibly interesting findings were beginning to emerge. The great thing about us being there with the data experts was that we could talk through their findings, how they found them, and how we could develop our understanding even deeper.
What we found out
Our data experts found a whole bunch of interesting statistics for us which we can’t wait to share, but for now my favourite findings for each questions are:
Q1 – What makes a fun and accessible activity?
Our Maths and English programme is designed to engage families in their children’s learning in a fun and accessible way, using hands-on, play-based activities. These activities have been categorised by our Content Team based on features such as whether they are indoors or outdoors, whether they are parent-led, or whether they require print-outs. We wanted to know if there were particular categories that lead to less engagement, particularly for disadvantaged families.
Some activity categories did lead to more engagement, and some led to less. For example, all groups of parents are significantly less like to do messy play activities, which is probably to be expected.
Intriguingly, some categories showed different effects for disadvantaged and non-disadvantaged families. For example, disadvantaged families are less likely to engage with activities that are parent-led, whereas their more affluent peers are more likely to choose to do these activities.
Q2 – What can we learn from the comments left by pupils and parents?
Every time a family records a reading entry or activity on our platform, they have the option to leave a comment and can choose whether to allow us to use that comment for research purposes. We have a bank of over half a million comments that we would love to be able to analyse in a meaningful way, to gain insights into how families found our activities as a whole.
The volunteers used Natural Language Programming (NLP) to gauge the sentiment of parent and child comments. From the sentiments they made a dashboard for our content team to easily see a summary of feedback across particular activities. For example, we found that some of our fractions activities had more negative feedback and featured sentiments such as “difficult” and “dad” more heavily.
Q3 – What features are driving engagement for our Maths and English activities?
We wanted to know whether there were certain markers outside of the activity categories that make families more likely to engage in our Maths and English activities. Particularly, is a family that uses our digital reading log more likely to engage in Maths and English activities?
We found that using the digital reading log significantly increases a family’s engagement with the Maths and English topics. It was amazing to see this in graphs and charts and it backs up what schools have already been telling us throughout the year!
Q4 – What features are driving engagement for our digital reading log?
Similar to question 3, are there certain markers that make families more likely to log reading? For example, how does the reading target affect whether the target is reached, or whether reading is logged at all in a week?
We found that the disadvantaged families using our digital reading log are more likely to keep logging reading with their child when a lower weekly reading target is set.
We look forward to sharing these insights with schools. A big thank you again to the DataKind team for these findings and for hosting such a rewarding weekend!
