The data comes from Mapping Museums project.
"The project’s research team has gathered, cleansed, and codified data relating to over 4000 UK museums - almost double the number of museums covered in any previous survey. It covers the period from 1960 to date."
Definitions of data points - https://museweb.dcs.bbk.ac.uk/glossary
Key Findings - https://museweb.dcs.bbk.ac.uk/findings
Let's load in the tools that we'll need for this EDA:
We may want to add data labels to some graphs so we need to write a script so we can call it when we want data labels:
Let's now read in the data:
I first of all want to have a look at the count of museums per country and whether or not they are still open:
Now let's plot the same but split the data by whether or not the museums are accredited. Accreditation is the professional benchmarking scheme for the UK. It is run as a UK partnership between Arts Council England, the Welsh Government, Museums Galleries Scotland and the Northern Ireland Museums Council. To gain accreditation museums must reach and maintain agreed standards of governance and organisation, collections care and management, public access and engagement with users. Museums must also meet the terms set by the Museum Association definition of 1998.
Let's split this a different way - this time, I want to see how the size of the museum affects whether it is accredited or not:
From this chart, we can see that most open museums are small in size. Additionally, it's more likely that unaccredited museums are small, and more likely that medium and large museums will be accredited.
Let's have a look at museums by Subject Matter:
I'm interested to see how this is split by country and I want to see this as a percentage of the total museums specific to each country:
We can see that Local Histories is the most popular type of museum in Northern Ireland (19.1%), Scotland (32.6%), and Wales (27.4%), but War & Conflict is the most populous in the Channel Islands (45.8%), and Transport is the most populous in Isle of Man (26.6%). War & Conflict is also quite a common subject matter in Northern Ireland (13.8%). These insights are probably closely connected to the history of these countries in the UK which may have led to an increase in interest past events and museums have been built to accommodate that.
Finally, let's finish with a fancy map since we have geospatial data to play with. I'm using the scatter_mapbox on plotly for this:
And that's it for this one! I've recently read Show Me the Numbers: Designing Tables and Graphs to Enlighten by Stephen Few and have utilised aspects of the book in my graphing for this EDA. In particular, the table showing the 'Subject Matter' breakdown is using many design aspects explained in the book which helps to better present your data.
Catch you all on the next EDA article! If there's a dataset that you'd like to suggest I look at, please feel free to drop me a message.
❤️ Enjoyed this article?
Forward to a friend and let them know where they can subscribe (hint: it's here).
