📈 Cleveland Museum of Art Collection Statistics

Taking a look into the open access data for the Cleveland Museum of Art.

View here - Code here

This is similar to the Metropolitan Museum of Art collection statistics I previously did, but has some notable differences:

1. Including the write-up inline with data
2. Different museum collections
3. Pre-processing is done entirely in the browser
4. Different charting libraries
5. Different chart types

Some opinionated choices made about how to do this:

• Since I’m pulling from a regularly-updated source, I want to do processing in code rather than a GUI tool
  • This makes it harder to handle free-text fields, but makes updating the data much more reliable
• I gave up trying to make the x-axis text for the first couple charts readable
  • Trying to go smaller resulted in unreadable text
  • Trying to go diagonal didn’t work out cleanly
• Make all functions outside of the specific charts as generic as possible
• Use the Cleveland Museum of Art website’s color scheme
  • I also didn’t feel that color or point radius were effective in these charts, especially the latter when points were close together

Notable edge cases:

• Some accession numbers follow an old format of ID.YYYY... instead of YYYY.ID...
• A few item didn’t have an accession number, (which is not supposed to happen…). I chose to ignore them

What was presented:

• Basic stats
  • Total items
  • Number of departments
  • Number of Types (akin to the Met’s Medium)
  • % of collections that are marked public domain
• Number of items per department
  • Gives information about how collections fit into the organization
• Number of items on display per department
  • Gives information about what the museum is presenting
  • When compared with the total number of items per department it can create interesting results
• Number of items by Type
  • Shows the kinds of items in the museum
• Number of types over number of items, by department
  • Searching for correlation between size of a department’s collection and the kinds of items it contains
  • Can tell us about how thorough a department is in utilizing metadata
• Percent public domain over count, by department
  • Of personal interest as someone engaged in open source
  • (Open access is also a museum value)
• Growth per year
  • How the conflation of a lot of factors influenced the size of the collections

Process:

1. Downloaded data
2. Examined in OpenRefine
3. Visualized more general info in d3
4. Deeper examination in OpenRefine
   • Used filters based on string manipulation to search for more complex data and prototype logic
5. Built more complex logic into summarization and charts

Things I didn’t get to but would like to look more into

• Per-department stats
  • I did this with the Met
  • How would it mesh well with narrative style?
• On-display per type
  • Just didn’t get to; already have the data summary for
• Culture field
  • Unlike the Met, this has better formating
  • Most things are Primary, more specific stuff, so can split on the comma with reasonably good accuracy
  • Would need to do a <1% aggregation again
• Estimated year of item
  • If start > end, then BC
  • If start === 0 && end === 0, then ignore
  • If no start nor end, then check text column; if /[0-9]{4}/ then single year
    • Couple similar cases (e.g. YYYY-yy for start/end years)
  • Notable edge case
    • Roses project (ongoing so no end date)