1. Tremendous diversity of skills
   
2. Ensuring good group chemistry
3. Limited time
4. Different goals?
5. Free-riding

1. Can take on many forms (from original empirical research to creation of open-source tools)
2. In one week, many groups will only find enough time to create a proposal, though some may have pilot results by the end of the week.
3. Limited seed funding may be available at your site for pilot research and/or data purchasing, cloud computing costs.
4. Additional funding may be available at your site after the end of the week.

1. We crowdsource a list of research interests in a google sheet (5 min)
2. Each person enters a “1” under their research interests (5 min)
3. We identify maximally similar clusters of participants
4. We identify maximally different clusters of participants

1. Maximally similar clusters come up with a group project (30 min).
   
2. Maximally different clusters come up with a group project (30 min).
   
3. We start a new google doc. Each person takes 5 minutes to write down their favorite group project idea.
   
4. We have a 20-30 minute discussion of ideas in the google doc with all SICSS participants at your site.
   
5. Everyone puts their name next to the group project they want to join
   

1. Have lunch today with your group in order to begin discussing your project.
   
2. It is ok to change groups until the end of today.

Monday afternoon: refine group project idea

Tuesday morning: write brief (<1 page) proposal

Tuesday afternoon: the organizer of your site will begin responding to seed funding requests on a rolling basis (if funds are available)

Tuesday afternoon-Thursday night: work on group projects

Friday (all day): group presentations (with feedback)

A document that contains a presentation of your group project that is between 10-20 minutes that explains:

1) Why your group project is important

2) What are your hypotheses?

3) What will you collect?

4) What are the next steps?

Note: not all projects will work; if yours fails, please write a post-mortem that explains why.

Create a new row with your name in the first column and a “1" under each of your research interests:

https://bit.ly/2WIrIYh

# Read in interests from google sheet
library(googlesheets)
SICSS_sheet <- gs_url("https://docs.google.com/spreadsheets/d/1RIeYYvtzEX2kkrXspP2npWBSd9_rRqtw6RywlGvCpsU/edit#gid=565698675") 

# Organizers, replace "YOUR SITE" with your site's sheet name below
interests <- gs_read(SICSS_sheet, 
                     ws="YOUR SITE")

library(dplyr)
# Create distance matrix
interest_mat <- interests %>%
  tibble::column_to_rownames("Name") %>% 
  as.matrix()
dmatrix <- dist(interest_mat, method = "euclidean") 
fit <- hclust(dmatrix, method="ward.D") 
# Display dendogram
plot(fit) 
# Draw dendogram with cut points for groups 
rect.hclust(fit, k=6, border="red")

# Identify all unique combinations of participants of size k (4)
all_combs <- caTools::combs(interests$Name, 4)
# Calculate mean distance within groups
dissimilar_groups <- as.data.frame(all_combs)
dissimilar_groups$score <- NA
for (i in 1:nrow(all_combs)) {
  temp_group <- interests[interests$Name%in%all_combs[i,],]
  temp_dist <- mean(dist(temp_group[,-1]))
  dissimilar_groups$score[i] <- temp_dist
  print(i)
}

# Retain groups with largest distances
dissimilar_groups <- dissimilar_groups %>% 
  arrange(desc(score))

# Use `max_diverse` to assign maximally diverse groups
max_diverse(dissimilar_groups)

Meet in maximally similar and dissimilar groups for 30 minutes. Site organizer will create google sheet where project ideas will be listed. At the end of each 30 minute period, one group representative should write the name of the project and a brief (less than three sentence description). After the end of the exercise, put your name next to the research project that you are most excited about joining.