So you want to help ADSEI build authentic data science into the curriculum? Here are some ways you can help!
If you have interesting content you would like to share with us, ideas for activities or blogs, or want to volunteer your time email us: contact@adsei.org.
Summary of Ways you can help
- Find interesting datasets, put them in a usable format, annotate them (details below)
- Create interactive visualisations of more complex datasets to allow kids to explore & understand them
- Create/find/write instructions for tools to help kids build their own visualisations (more creative than just graphs!)
- Write up cool data science stories in simple language: ways it has failed, amazing things it has made possible
- Make short videos explaining data science concepts
- Write interesting data science lesson plans/project ideas for collecting their own data
- Spread the word!
If you have a cool dataset, we need it to have:
- Written explanations (or videos)
- What is the data?
- How was it collected?
- What does each field mean?
- What are the limitations of the data?
- What can the data tell us? (and what CAN’T it tell us?)
- Easy format (usually csv), clearly labelled, simplified if necessary Where csv isn’t possible, software to visualise/manipulate the data (or instructions on using eg nationalMap or MyMap or other systems)
Data must be open source and shareable on the ADSEI website.
Other Helpful Resources you can create:
- Project ideas (more detail under “Classroom Activities below”)
- Working with real datasets to solve actual problems
- Or collecting & using data to tackle local community problems
- Visualisations, Written or video discussions of:
- Good graphs/Visualisatons (and why they work)
- Bad graphs/Visualisations (what’s wrong with them, and how they can be made better)
- Data Science examples – Write ups or short video explanations of
- Useful things you’ve done with data science
- Cautionary tales – especially where initial analysis is misleading
- Interactive visualisations (such as this awesome example from Eliiza) that help kids explore and understand more complex datasets
- Python library for easy visualisations
There are huge numbers of open source datasets available. Finding them is easy. But you can’t expect your average teacher to (a) find a relevant dataset (b) make sense of it and (c) build lessons around it. They don’t have the time, or often the skills. The more help we can give them, and the easier we can make it, the more likely they are to adopt it in the classroom.
Teachers are really daunted by Data Science – they think it’s all building Machine Learning systems in Tensorflow, and waaaay out of their scope. So it’s really important we bring them in gently, building on skills they already have.
Classroom Activities you might write up include:
- Exploring what the dataset means – what it can tell you, what it can’t, what its advantages and limitations are
- Analysing the dataset in different ways – especially exploring the difference between different ways of looking at the data and how they might mislead you. For example, the renewables dataset – simply sorting it tells you the top postcodes, but it’s easy to infer that since the top postcodes are nearly all in WA and QLD, they are the top states. When you calculate the average, though, they turn out to be really low in the rankings. It’s great to write up (or video) discussions of this kind of thing so that teachers know how to run the discussion.
- Exploring the best kind of visualisation for a particular dataset – or more accurately, for a particular question for a particular dataset. Examples of this, and even video explanations, would be awesome.
- Some Sample project plans (note that I can polish and add curriculum links etc if needed, just the annotated datasets and some analysis/exploration ideas would be super useful):
Biodiversity for Primary Schools (collecting own data)
Year 5/6 Renewables project (collecting own data)
Year 7/8 Renewables Project (using online dataset)
We have a page of data repositories if you need help finding interesting data (let us know if you know others we should add!)
Some datasets will need to be simplified for school use. For example, the Happiness report dataset has been broken up into by-year and by-country data, from the original which contains hundreds of countries and ~10 years’ of data.
The aim for now is that most projects should be doable in a spreadsheet, as we don’t have enough teachers with Python/coding skills to teach kids to do that. But don’t be afraid to include extension activities that need code!
Dataset Requirements:
- CSV format for preference
- Small enough to open in Excel (ideally varying in size, smaller datasets of a few hundred lines for beginners through to several thousands for more advanced students)
- Consider sampling or top and tail depending on the dataset
- Or breaking up by relevant categories (for example voting data I could have broken up by electorate)
- Limit the number of columns to just what is needed for project & understanding if there are squillions (ideally columns shouldn’t scroll too far off the screen, unless they’re really important to the project).
- Provide raw dataset & simplified dataset if trimming (so kids who can code can use the full one)
- Don’t clean unless it’s insanely messy, because cleaning is a great lesson – So dates in different formats etc, leave them be!
- And please provide clear plain English explanations of what the data is, how it was collected, what the units are, what each field means, etc. (in a separate google doc)
- If they are datasets that change over time, or provide timely snapshots of information, please provide links to the originals, so that we can grab up to date versions for projects.