Meeting Minutes Jan 30 18

2018-01-30

Location	Time	Duration
CSC-250	10:00 - 12:00	2hr

General Discussion

• Eleni has spoken: NLTK (implies python), knowledge graphs can be ignored for now. N-gram extraction is a must.
• PDF scraping must first cut up the PDF's by attachment, then associate the attached PDF sections to their topic

Work divided into 3 sections for purpose of the Gantt chart

• Scraping + Databases
• NLP
• Front-end

Brainstorm visulizations

• Timeline (https://github.com/hkelly93/d3-relationshipgraph , except we envision a left to right layout)
• Results (following timeline, filter by best results, filter by frequency)
• Wikistyle graphs, committee and who is on it, person, and committees they are on
• Nightmode
• Loading graphics (magnifying glass over documents)
• Sankey diagram (show which committee a person spends most of their time)
• Show which percentage of meetings have been attended by which percentage of members (all meetings attended by member A, 50% by member B)
• Show which

Action Decisions

• We will email Ann by tomorrow morning
• We will finish all Sprint 1 issues by Wednesday night 11:59pm

Technology Decisions

• Python for back bone
• NLTK for NLP
• Elasticsearch for search
• pdfquery for PDF scraping
• Bootstrap for HTML/CSS/JQuery
• d3.js for visulizations
• Django for web framework (will check out flask but that is a fallback)