Presented by Sahar Alhassan, Kaitlyn Lynes,
Deepa Mehta, & Ilya Perepelitsa
December 14, 2015

• National Science Foundation (NSF) is a key engine that funds US-based R&D
• Our group focused on NSF grant dollars allocated to US institutions from 1999 to 2015
• We focused on the relationships between grant type (new, continuing, revised) over time
• We also examined the types of approved grant projects, as well as institutions and disciplines that received federal funding
• We also began to explore geographic concentration of grant funding

• Started with the NSF Grant Database
• Collated yearly data into a separate subsets of a dataframe
  
• Aggregated data by year using loops and packaged
• Developed a series of plots

load("NSF.RData") 
library(ggplot2) 
newdata <- schoolsall[which(schoolsall$starting_date < '2015-01-01' & schoolsall$starting_date > '2014-01-01'), ] 
newdata1 <- schoolsall[which(schoolsall$starting_date < '2015-01-01' & schoolsall$starting_date > '2013-01-01'), ] 
newdata2 <- schoolsall[which(schoolsall$starting_date < '2015-01-01' & schoolsall$starting_date > '2012-01-01'), ] 
newdata3 <- schoolsall[which(schoolsall$starting_date < '2015-01-01' & schoolsall$starting_date > '2008-01-01'), ] 
newdata4 <- schoolsall[which(schoolsall$starting_date < '2015-01-01' & schoolsall$starting_date > '1999-01-01'), ] 
newdata5 <- schoolsall[which(schoolsall$starting_date < '2015-01-01' & schoolsall$starting_date > '1999-01-01'), ] 

• One year snapshot
• Continued and New Projects Received More Grant Allocation than Revised Projects

• We decided to look at this over two years
• This showed us a trend

• We looked at this trend over 16 years
• Towards the end of every year, we see more grant dollars allocated

H1: Less expensive grants are more likely to be granted at the beginning of the year, while more expensive grants are more likely to be granted at the end of the year. This reflects difficulty.

H2: Pressure to spend end of fiscal year Congress Appropriations results in increased demand for grants. NSF more likely to choose expensive grants to appropriate money more quickly.

• We broke it this trend down by action type
• Revised grants saw a large gap between 2001 and 2004

• During this period, most grants were in the Engineering, Mathematics and Geosciences
• Institutions appear to receive most of the grants directly

• Are projects with a longer duration allocated larger grants?
• Newer grants appear to be shorter in duration

• Revised grants are longer than continuing grants

• What accounts for the gap in data in 2004?
• Is this a gap in data or was research funding halted due to a government shutdown in 2004?
  

• Bimodal distribution in revised grants
• New grants have higher allocations

• More density among grants for STEM fields
• High density of grant allocation to polar programs

*Appears to be uniform, although some variation in Small Business recipients

• Grants appear to be concentrated in certain recipient types for revised grants (i.e. State Government)
• We see narrow concentrations spiking for Small Business Grant recipients

• Trends appear to be consistenet in all states and territories
• Territories appear to have more new grants within this period

• New grants in relation to continuing grants
  

• CA, DC, MA, and NY have the most grants
• CA received 15% of the grants (consistently)

• Improving the accessibility of this data
• Prepare the dataframe in a way that allows us to generate reports by research type, institution type, state, and key words in grant descriptions
• Be able to expand this data to other governmental datasets