[sims] SIMS 8 Press Release 30th March 2017

[bobsmith]

Usually because of lower pay - as @localzuk says, as things change and trusts/groups can afford to pay more for centralised roles, so the skillsets should increase too.

 

I would not consider it at all unreasonable to expect a Data Manager on a £35-40k salary to know T-SQL, SSRS, etc.

 

I'd love to see these job adverts for £35 to 40K!!!

 

The typical data managers I've met are office administrators with a jumped up job title on 20K max!

 

I was considered over paid on £35K and I've not met a lot of DMs on that salary (even in london).

 

I convinced my last school to send me on 2 microsoft SQL courses but I'd be amazed if I ever managed that feat again.

 

That said a properly qualified DM with that skillset could command a lot more cash in the private sector. I had a IT network manager leave and triple his salary doing much the same job for a private company.

[vikpaw]

Subscribe to the educational jobs forum. I frequently see jobs that are offering 30k+

Not necessarily data manager but some are.

Groups are certainly seeing the benefit of these roles.

We’ll always struggle with single schools seeing the value in a Data Manager, many still make the IT guy where that hat.

The value is often only seen when the incumbent de facto DM leaves and needs replacing.

[FishCustard]

That said a properly qualified DM with that skillset could command a lot more cash in the private sector. I had a IT network manager leave and triple his salary doing much the same job for a private company.

 

While I don't doubt that in general salaries are higher in the private sector, I'd likewise like to see how an IT manager could triple their salary by going private!

[elsiegee40]

I’d say £25-28k is average out of London, BUT that gets pro rated down as the jobs are usually term time plus 2 weeks.

 

More and more MATs and federations are advertising group posts at higher salaries

[matt40k]

Wow this has drifted off topic. Nice to know even the EduGeek mods are bored of talking about the vapourware.

[CloudGeek]

Wow this has drifted off topic. Nice to know even the EduGeek mods are bored of talking about the vapourware.

 

👍👍

 

Well put @matt40k

[NotSoFast]

Since we're already off-topic... Putting aside the ethical issues for a moment, I'd love to see what could be done by applying some AI to the huge amount of data in MIS systems, especially at trust level. Rather than humans trying to spot patterns and issues perhaps the system could flag things up for us? What are the early characteristics of a student who goes on to be a school refuser, or get permanently excluded (beyond the obvious!)? Who is skiving particular subjects? Which teachers' predictions look odd compared to others' for the same students?

[matt40k]

That isn't AI. That's more machine learning.

 

Imagine doing it at an Local Authority level (so hundreds of schools), then add police, youth offending, health data and other data sources. The problems isn't getting data or even analysing. It's getting permissions\policies sorted and turning it into actionable steps.

[Ditto]

Well, now you're talking a very interesting subject, "predictive analytics" - probably well off topic but I didn't start it! Before I hit the education sector, my last role was leading a development team working on Text Analytics - basically taking unstructured text and pulling out key concepts/sentiments for analytics with more traditional structured data. Whether predictive models can be built, say from written behavioural reports, only a trial would prove. But, I'm sure considering all the data we collect in schools could be put to better use. One model in the telecomms world was modelling 'churn' for phone contracts. The models basically identified those most likely to leave/join. You could then target your efforts in a much more focused way and save a lot of mis-placed marketing spend. One of my favourite elevator pitches was a model that predicted one of three sites on at a particular day/time/venue for a criminal to make his next robbery. The cops staked-out the top three and nabbed the baddy! For sure I believe similar analytics could be used to identify and target say intervention strategies. I'd love to see a trail done, but I just can't see it in any of the schools I've worked at as it requires a high degree of investment. Often it would be very difficult to get SLT to buy-in to the idea that a computer algorithm could spot trends/patterns that, given the ratio of staff/pupils, wouldn't have been spotted by the non-artificial intelligence!

[Ditto]

That isn't AI. That's more machine learning.

Imagine doing it at an Local Authority level (so hundreds of schools), then add police, youth offending, health data and other data sources. The problems isn't getting data or even analysing. It's getting permissions\policies sorted and turning it into actionable steps.

The key is often finding out which items of data are really useful to the model. Some of the best models I saw surprisingly ended up with a handful of data points. You right about the challenge of using that data and with the GDPR rules in this area, you'd need to proceed very carefully.

[NotSoFast]

That isn't AI. That's more machine learning.

 

Imagine doing it at an Local Authority level (so hundreds of schools), then add police, youth offending, health data and other data sources. The problems isn't getting data or even analysing. It's getting permissions\policies sorted and turning it into actionable steps.

Yeah, ok. I started with ML and swapped it. [emoji3]

[NotSoFast]

Yeah, ok. I started with ML and swapped it. [emoji3]

I was thinking about Capita or a big trust, (or even the DfE) making the investment in development not an individual school.

[PhilNeal]

LAs already have tools that can create risk profiles and flag up children potentially at risk. Studies have already identified risk factors that can be weighted. For example - moving house more than x times, high absence rates & parent with a court order. It's not AI and is only an indicator.

[matt40k]

Yup...

turning it into actionable steps.

 

The last point being the hardest.

[NotSoFast]

LAs already have tools that can create risk profiles and flag up children potentially at risk. Studies have already identified risk factors that can be weighted. For example - moving house more than x times, high absence rates & parent with a court order. It's not AI and is only an indicator.

Yes, we get this list of "RONI"s from the LA, but it's really telling us they are at risk AFTER they have been excluded/moved house/ been taken into care/ given an EHCP. I'm more interested in what these kids' profiles were like BEFORE a flag appeared.

 

I suspect attendance patterns and relatively low level behaviour issues would be signals worth looking at, along with other things (for example attendance at parents' evenings), but I'd assume the system itself would work out the things to look for.

 

Further off topic, I'd like a voice interface please. "Alexa, ask SIMS which lesson Joe Bloggs is in at the moment. / What were Joe Bloggs' SATS results?" etc etc. Again, plenty of data protection issues there, but still...

[matt40k]

Part of the problem is how .... umm how do I put this politely.... how out-of-date SIMS is, it's pretty problematic getting data sucked up into the Local Authorities system. It isn't cheap either.

[bobsmith]

some of our SLT already have a voice interface for SIMS - at least it seems to work that way for them.....

[DODICT]

Other MIS products have natural language querying - so using OS speech to speech this is possible now . . .

 

I'm not convinced its killer functionality just leveraging tech that is readily available but we came up with a few examples where it might be useful and takes away some barriers to getting raw data on screen easily.

[bobsmith]

Other MIS products have natural language querying - so using OS speech to speech this is possible now . . .

 

I'm not convinced its killer functionality just leveraging tech that is readily available but we came up with a few examples where it might be useful and takes away some barriers to getting raw data on screen easily.

 

but you can just pick up a phone and ask an admin person to look it up and read it out loud to you....?

 

ok /sarcasm mode off

[Banjo]

Yes, we get this list of "RONI"s from the LA, but it's really telling us they are at risk AFTER they have been excluded/moved house/ been taken into care/ given an EHCP. I'm more interested in what these kids' profiles were like BEFORE a flag appeared.

 

I suspect attendance patterns and relatively low level behaviour issues would be signals worth looking at, along with other things (for example attendance at parents' evenings), but I'd assume the system itself would work out the things to look for.

 

Further off topic, I'd like a voice interface please. "Alexa, ask SIMS which lesson Joe Bloggs is in at the moment. / What were Joe Bloggs' SATS results?" etc etc. Again, plenty of data protection issues there, but still...

 

The issue LA's have is that one of the biggest indicators for NEET is exclusion data/flags. When you profile NEETs you will pretty much find a history of FT/Perm exclusions. If you remove exclusion data from your modelling your identification of potential NEETs falls off the cliff. What you can do is profile the 'risk of exclusions' which does factor in low attendance and academic performance. Some LA's in London are moving to these models to predict issues within the Social Care arena.

[supportman]

Maybe a hint on the delays of Sims 8 and whats going on at Capita from their recent stockmarket update:

 

The unit was once a hodge-podge of 29 "siloed" businesses – that's what happens when you grow via acquisition – but unspecified "structural and operational changes" had addressed this, Capita claimed. It has also built a "development base" in India to offshore application development "at a lower cost base".

 

SOURCE: https://www.theregister.co.uk/2019/03/14/capita/

 

So basically they have outsourced their software development to India to save costs. And maybe that's why we haven't seen any move to Software as a Service in the last 5 years? I don't know many Indian agile software companies so hardly surprising.