Dear Data to Insight colleagues –
Please find below another bundle of news, within which I hope you’ll find something of interest. As always, if you want to talk about this stuff, or have ideas for how we can do useful things – or how you’d like to get involved – then do get in touch and let me know what you’re thinking about.
ChAT updates – bug fixes, weekly charts, and file size reduction
Quality CSC Data – collaborative coding project kick-off
D2I code library – python learning resources now available, including example “drift” tool
D2I “open house” – informal show and tell sessions soon
Website accessibility
Credits
1. ChAT updates – bug fixes, weekly charts, and file size reduction
We’ve updated ChAT and you should be able to find the latest versions at the above links.
Alongside several small fixes which you can check in the version log, this time we’ve added a new “ChAT_Weekly” tab to the ChAT which replaces the previous standalone “COVID ChAT” workbook in providing weekly activity trend analysis for those lists where that feels appropriate – mostly front door activity. We’ll be retiring the “COVID ChAT” as a result of this, to save us time updating the same hidden tabs in so many separate files, and hopefully the switch means more people will get some value out of those weekly visualisations.
You may also notice that John has made some changes to ChAT behind the scenes which have taken the file size down from 42MB to around 30MB. There’s more we can do on this but we’re doing a bit at a time so that, if something goes wrong, we know which bit to put back how it was! Any feedback, just let us know.
We await news on updated adoption scorecards data, which we had expected soon, and mid-year population estimates which are this year expected in November. Depending on the timings, we may wait to combine this with the December DfE data publications – more on this later.
2. Quality CSC Data – collaborative coding project kick-off
September 8th at 10:30 – session A
September 13th at 13:00 – session B
The Quality CSC Data project team, with help from colleagues at Wigan Council who are hosting the project, have been hard at work setting up to repeat our successful 903 data validation project, this time looking at CIN census data. There are improvements to how the tool will work, and improvements to how the project will run for collaborators. The first output will be a web-based data validation tool which you can use year-round to validate CIN census data, speeding up data quality work, easing the burden at year-end, and improving data quality in general. We hope to launch the first version of this around November time.
The project itself will be a joint development project where you can sign up to learn some Python and code some of the tool yourself, with help from our expert programmers and learning materials. This is a great opportunity to learn a new technical skill working on data you know well and with help on hand. It’s also a great way to demonstrate how useful these approaches can be, if you need evidence to justify expanding your toolkit locally. If you’re at all interested, please contact Will and ask for an invitation to one of the following kick-off sessions:
If you can’t make these dates, let Will know and we’ll figure something out.
3. D2I code library – python learning resources now available, including example “drift” tool
Will has been working on learning resources for the Quality CSC Data project, and as part of that he’s developed a series of Python exemplar tools and learning workbooks focused on children’s services data and D2I Excel tool functionality, which are intended for beginner programmers to use as learning aids. Will writes:
They come in three formats: usability, development, and learning versions. Usability versions are the easiest to use, whilst the development versions are ideal for those with Python experience who want to practice tinkering with or customising the tools, which is also great practice for learning how to work collaboratively on code. Finally, the learning versions of the tools are worksheets which you can work through to “fill in the gaps” and refer to complete “answer” workbooks when you need.
The goal of the usability and dev versions is to show off the kind of functionality Python can have with a little experience, and the learning versions are to give you a head start in making your own tools. There’s also a file full of useful reference examples of particular tasks you’re likely to need to do as a children’s services data analyst working in Python.
As a D2I site member you can access the guides at the link above. To get access to the D2I code library on GitHub, or any help with these resources, please email William.Levack-Payne@eastsussex.gov.uk
As part of that Python work, and on the same page of the website, you’ll find a link to a new “Drift” tool which we worked on as an example of a project we could easily develop in Python. This is a bit of a departure from our usual approach but we hope someone will find it useful, either as part of the Python learning process or as an analysis tool in its own right. To use it, you’ll need to be able to work with Jupyter notebooks, and you’ll need to generate long data extracts of referrals, CP starts, and CLA starts, each containing a Child ID and a start date. The tool will then calculate things like the varying duration between referral and plan start, to indicate whether children are tending to take a longer or shorter time now than in the past to arrive at the support plan they eventually need.
I first came across the idea of “drift” analysis in the middle of an Ofsted inspection and pulled together my analysis in a panic using Microsoft Access. It showed something we hadn’t noticed at the time, and became something we referred back to quite often. If there’s wider interest, we can look at the possibility of an Excel version, though I suspect it won’t be as quick to make as was this Python notebook!
4. D2I “open house” – informal show and tell sessions soon
Wednesday 21st September at 13:00 (and every two months thereafter)
The core D2I team recently met for our first proper team meeting, and one of the outputs of that meeting was that we should be hosting more show-and-tell sessions to keep each other informed of what we’re working on – clever things we’ve figured out, useful approaches for particular tasks, common problems we were wrestling with, and so on.
It seems a huge shame to keep that kind of thing to ourselves, so, in the spirit of “working in the open”, we’re going to host these as “open house” sessions, where anyone can drop in and listen and contribute if they’d like. We’d be very happy to have people bring along things they’re working on or interested in to talk about. The first one is on Wednesday 21st September, and we’re planning to run them on the third Wednesday of every second month thereafter. If that’s sounding hard to remember, email John and he will happily send you a recurring invite – just drop in as and when suits you. If nobody comes the D2I team will talk among ourselves, and if 100 people come then we’ll panic think about running them more frequently.
5. Website accessibility
You may notice some small changes to the D2I website layout in recent weeks; these are in service of recent accessibility work on the site. Will has worked through a long report checking how well the site performs from an accessibility perspective and has made several such changes. In the footer of the site, you’ll find the first version of our new accessibility statement if you’d care to read more (link above).
The key point to note here is that, because our site serves a relatively small core community of users, we think we can also afford to cater to individual accessibility requirements to make sure nobody’s left out of what we’re doing. If for any reason our website isn’t working for you, and you’d like to suggest a change or ask us for help in accessing anything we provide, please do get in touch, either via the site or by reply email to me, and we can talk about how we can better help.
6. Credits
Credit for recent improvements to our data tools is due to colleagues in:
Leeds
East Sussex
Islington
Medway
North East regional group
Dorset
Haringey
Knowsley
Hounslow
Bristol
Liverpool
Waltham Forest
Bournemouth, Christchurch and Poole
Somerset
Hertfordshire
Yorkshire and Humber regional group
Camden
Bury
Ofsted
Kirklees
Calderdale
North Yorkshire
Luton
Wigan
Department for Levelling Up, Housing and Communities
Social Finance
That’s it! If you have any comments, queries or ideas that you want to share, just let me know.