Content

[vaidap]

Homepage (editions)

• [ ] number of words in edition + number of volumes
• [ ] longest word defined + definition
• [ ] most common words (wordcloud) NO
• [ ] 5 most common words (excl. stopwords)?
• [ ] graph of topics referenced (make long and scrollable)
• [ ] suffix/prefix? longest word related to x topic?
• [ ] definitions with most refs
• [ ] longest/shortest definition?

Data page (description)

• [ ] Counts of references

Ideas for "main course"

• [ ] connected cloud (see if can do manually for small scale, for definitions with most refs)
• [ ] topic presentation (see words referencing topics & their definitions)
• [ ] research case studies, technology we could use

[vaidap]

Questions, messages? What do you want to ask the data and what kind of discoveries have you made? Focus on communication. What are you saying? Show interesting things

• how topic popularity changes over time, proportional change over the years for topics (line chart)
• term across time: does it saturate or go up all the time?
• timeline of events during encyclopaedia (colonialism, etc): how world events influence encyclopaedia topics. highlight key events

https://www.gapminder.org/tools/

[vaidap]

Alexander feedback

• Program: functions to recycle, data cleaning module, commits/PRs. structure
• Aggregate constructed data that end user sees -> they don't see the messy content behind it
• 1000 word report and website, plus a video with what we want to highlight
• summary stats
• edition + time: line chart of growth/decrease in topic popularity
• who is this for?

[vaidap]

• presence of words in each edition? hover to see words
• word embeddings: should be straightforward, do on small section of data & change against time. meaning at that time: stability in meaning https://arxiv.org/pdf/1605.09096.pdf

[vaidap]

https://app.slack.com/client/TPHHYS55M/CPK3E7KLP https://arxiv.org/pdf/1605.09096.pdf

[vaidap]

• send alexander a list at beginning of next week: what we will get done for sure, what is less likely
• github activity: put all code on there & get github part of cw done
• distribute tasks
• clean data up a bit more?
• process data in a way that's easily exported & don't need to redo each time

[vaidap]

what to use: http://www.tfidf.com/ or just word frequency: TF(t) = (Number of times term t appears in a document) / (Total number of terms in the document).?

[vaidap]

anatomy medicine architecture agriculture botany

-> frequency in each edition -> words related to: all defs with "see Anatomy in it"

-> counts of how many different references there are, and graph of how many each referenced

[vaidap]

• total referenced topics for all editions: bar chart OK
• 5 topics: bar chart for how important, proportional OK
• make code for slider (with longest word & def, most ref word & def, highlight, volumes) OK

[vaidap]

• changes of tops across the editions: podium
• pick one word in each of the top 5 topics: to see how definition evolved across editions
• slides: what are the findings: topic popularity, interesting new topics, how language was used, 7th&8th edition changes (less cross-referencing, also big topic like anatomy has less in 7th&8th, and more longer-article subheaders are being picked up, and some words just don't have any entries in final two editions), general structure of editions/data, data we collected. mention that less popular might mean that it was merged into a long-form article with all the entries in it rather than referencing the topic from other places, so actually was upgraded. this just investigates cross-referencing. vols don't include supplements, etc
• provide Ava with picture OK
• description of our project OK
• reply to Sarah OK

[vaidap]

for submission:

• report
• video?
• github assignment
• clean up github & post a public-facing version on website

[avahsieha]

for website adjustment

• add our Linkedin links to the profile page
• add axes to the popularity of the topics
• remove the "kid" word of the game
• screen recording for the website