Data Thing Part 1: What shall we do and why?
This post is the first in a series about the project currently known as Data Thing.
We begin at the stage where we have no idea what we are going to do, but we know we need to do something.
I learn things by doing them, so join me as I muddle my way through working out what this project is by trying stuff.
In this case WE usually refers to ICT, and USERS are staff in business areas.
We begin with a priority in the ICT service plan to support the Council to make better use of data...
The situation
- The council wants to be data driven and do all that cool sounding analytics stuff the kids are into these days.
- Use of data is currently focussed on dashboards and reports that serve no operational purpose beyond being updated and looked at.
- Our data is messy and incomplete. It is stored in a hodge podge of locations including:
- On premise (data centre) Oracle and SQL databases
- Supplier hosted databases accessed by VPN, API or scheduled extract
- Supplier maintained databases e.g. Redshift
- Excel and CSV files
- Many PDFs 🥴
- IDOX DMS
- People's heads
- Word documents
The challenges
- There is no clear goal beyond ✨be data driven and use data to inform service delivery✨
- There is no clear definition of "our data". Is it the data we own or the data we find useful or both or something else?
- The budget for the project is currently £0. Any spend will require a business case for approval.
- Technical skills are inconsistent across the organisation, but there are pockets of good practice and people who are willing and excited to learn.
- Technical staff spend a lot of time working on ad hoc SQL reports for limited, very specific use cases. There is a backlog of requests which we cannot keep up with. Its not a very efficient way of working.
- Legacy systems often have database structures that make no sense to end users or normal people. Giving users direct access to create their own reports will require more support than we can offer.
So what now?
Now comes the bit where I try to balance spending time planning and pondering, vs getting stuck in with the work. Based on the above we decide:
- We should look at a way we can bring our random datasets together and give users access to them in a way that is easy for them to make use of and has appropriate access controls.
- It needs to be cheap, so let's look at tech we currently pay for that we could make use of.
- Users would benefit from technical training but also some support on identifying and evidencing the benefits of data projects.
- We've found success in the past by starting up a little demo project with colleagues who are keen and interested, and then presenting this to senior management and other teams to generate buy in and momentum. So we'll take this approach over starting with a top down strategy.
Other posts with tag "data thing":