Scope of the possible with Power BI

Previous attendees have said…

• 17 previous attendees have left feedback
• 100% would recommend this session to a colleague
• 88% said that this session was pitched correctly

Three random comments from previous attendees

• Fab intro into what Power BI can be used for. Brendan explained everything really well.
• Was interested in whether Power BI might be a useful way to provide access to our ‘siloed’ data to management. The answer is yes, if only NHS restrictions would allow us to use Power BI 😂
• It was great to get a more detailed look at what the possible limitations are when considering Power BI

Session outline

• this session is a non-technical overview designed for service leads
• Why Power BI, and why this session?
• Power BI demo - build a simple dashboard
• Strengths and weaknesses
  • obvious
  • less obvious
• Alternatives
• Skill development

Power BI?

• newish (c.2015)
• based on Microsoft’s SQL reporting products
  • proprietary, closed-source, paid-for
• integrates functions from several other Microsoft data products (bits of Excel, PowerPivot, PowerQuery…)
• produces interactive dashboards

Why this session?

• there is a lot of fluff talked about Power BI
  • e.g. elaborate visualisations, fancy real-time data products…
• Power BI in general is promoted as a no-code tool
  • that’s just not true, as we’ll see
• Power BI has a sales-y focus
  • that conceals some of its most useful properties
• so Power BI, as a platform, requires some translation to the complicated reality of health and social care data
  • and Power BI offers real strengths in managing and exploiting our data, even when the pain-points are taken into account

Power BI demo

We’ll use a pair of Excel files. These are based on three datasets from the Scottish Health and Social Care Open Data portal:

• GP practice size data - which is based on the GP practice details dataset and the Health Board 2014 - 2019 dataset
• Demographic data - which is based on the GP practice populations dataset

GP practice data

Name	PracticeListSize	Postcode	HBName	GPCluster	Code
Bangholm Medical Centre	11189	EH5 3AH	NHS Lothian	Edinburgh North West - Bridge	70272
The George Practice	1651	ML7 5DS	NHS Lanarkshire	Newmains, Shotts & Harthill	62897
Skerryvore Practice	8643	KW15 1NZ	NHS Orkney	Orkney	38012

Load some data

Lots of different data loading options

Preview

Friendly tools for previewing data

Add to a map

Clever use of commercial postcode data

Add interactions

Visualisations are interactive

Publish

Add more visuals

30-odd built in, hundreds of free additional visuals, thousands of (often dodgy) commercial add-ins

Add more data

• we could add the health board names, to make our visual more useful
• we could also get GP practice demographics

Add more data

Totally different data, but harmonised process for loading

Re-shape that data

Neatly-integrated PowerQuery, gives loads of scope for tidying and fixing data

Data modelling tools

Full suite of relational tools for the more ambitious projects

Pre-packed visuals

Microsoft’s Tornado chart

Demographics

These can be cross filtered to give national/board/practice level insights

Strengths

• by far the easiest way of producing interactive data products
• great tools for tidying data
• good for iterative projects
• data hubbing / self-service data
• scaling
• potential to manage complex and sensitive data on existing infrastructure

Weaknesses

• really needs clean and tidy data
• publication can be complicated and expensive
• steepening pain curve. Easy to start projects, but more involved analysis is messy
• complex IG landscape
• messy skills development journey