Add a README with stuff we brainstormed about this data.

README.md

@@ -0,0 +1,43 @@
+# Participants
+
+- @jamey\_sharp
+- @stagedtheology
+- @tklee
+- @galgeek
+- @edrex
+
+# Raw Data
+
+...is what this repository contains.
+
+Also here's the proposal json converted to a spreadsheet:
+<https://docs.google.com/spreadsheets/d/1XZfB3b9_TfQ855yhBCEFHxtTeI9AX00rvj74py-vRp0/edit?usp=sharing>
+
+# Questions
+
+- Can we find like-minded people based on favorites?
+- Can we generate BOF sessions in advance?
+- Can we design a schedule to minimize conflicts among sessions?
+  - Cost assignment (based on timestamp?)
+  - HARD CONFLICTS: Speaker can't appear in multiple locations
+- Can we optimize room size based on favorites?
+- Do favorites predict attendance?
+- Do clickthroughs on description predict attendance? (We don't have data, but maybe!)
+- Can we (from full text description) predict:
+  - level?
+  - attendance?
+  - \# of favorites?
+  - track?
+  - which talks will be accepted?
+
+# Tools
+
+- cplex
+- linear programming: AMPL
+- <http://www.neos-server.org/neos/solvers/> \<- web interface to solve LP/NLP/MILP directly
+- ibm ilog studio
+- coin or (repository)
+- informs
+- SMT solvers (e.g., solve sudoku puzzle based on initial values)
+- Bayes' theorem!
+- Tabu search: <https://en.wikipedia.org/wiki/Tabu_search>