Setup
Build the Table
- Write to File
Diagnostics
- Missing or Extra XPATHS
- Failes URLs
XML Table Structure
- Create Network Edgelists
- Visualize Tree Structure
XPATH Summary

Setup

Load Packages

library( irs990efile )  # functions for parsing efile XMLs
library( dplyr )        # data wrangling
library( purrr )        # data wrangling 
library( pander )       # formatting
library( knitr )        # formatting
library( DT )           # table printing
library( httr )         # web requests 
library( xmltools )     # xml utilities
library( xml2 )         # xml utilities
library( XML )          # xml utilities 
library( data.tree )    # network visualization
library( networkD3 )    # network visualization
library( igraph )       # network visualization

source( "R/rdb-functions-v2.R")
source( "R/utils.R")

Data

head( index )

The Concordance File

The concordance is an RDB to XML Crosswalk that maps all xpaths onto a specified set of variables. It is located in the irs990efile package.

The full data dictionary is available on GitHub.

DATA DICTIONARY

concordance %>% 
  filter( rdb_table == table.name )

xpath	id
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/AmountOfCashGrant	1
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/AmountOfNonCashAssistance	2
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/DescriptionOfNonCashAssistance	3
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/MannerOfCashDisbursement	4
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/MethodOfValuation	5
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/PurposeOfGrant	6
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/Region	7
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/CashGrantAmt	8
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/DescriptionOfNonCashAsstTxt	9
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/MannerOfCashDisbursementTxt	10
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/NonCashAssistanceAmt	11
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/PurposeOfGrantTxt	12
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/RegionTxt	13
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/ValuationMethodUsedDesc	14
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/AmountOfCashGrant	15
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/AmountOfNonCashAssistance	16
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/DescriptionOfNonCashAssistance	17
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/MannerOfCashDisbursement	18
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/MethodOfValuation	19
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/PurposeOfGrant	20
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/Region	21

xpaths <- as.character( t.xpaths$xpath )
el <- create_edgelist_v1( xpaths )
nd <- FromDataFrameNetwork( network=el )
print( nd )

##                                             levelName
## 1  Return                                            
## 2   °--ReturnData                                    
## 3       °--IRS990ScheduleF                           
## 4           ¦--Form990ScheduleFPartII                
## 5           ¦   °--GrantsToOrgsOutsideUS             
## 6           ¦       ¦--AmountOfCashGrant             
## 7           ¦       ¦--AmountOfNonCashAssistance     
## 8           ¦       ¦--DescriptionOfNonCashAssistance
## 9           ¦       ¦--MannerOfCashDisbursement      
## 10          ¦       ¦--MethodOfValuation             
## 11          ¦       ¦--PurposeOfGrant                
## 12          ¦       °--Region                        
## 13          ¦--GrantsToOrgOutsideUSGrp               
## 14          ¦   ¦--CashGrantAmt                      
## 15          ¦   ¦--DescriptionOfNonCashAsstTxt       
## 16          ¦   ¦--MannerOfCashDisbursementTxt       
## 17          ¦   ¦--NonCashAssistanceAmt              
## 18          ¦   ¦--PurposeOfGrantTxt                 
## 19          ¦   ¦--RegionTxt                         
## 20          ¦   °--ValuationMethodUsedDesc           
## 21          °--GrantsToOrgsOutsideUS                 
## 22              ¦--AmountOfCashGrant                 
## 23              ¦--AmountOfNonCashAssistance         
## 24              ¦--DescriptionOfNonCashAssistance    
## 25              ¦--MannerOfCashDisbursement          
## 26              ¦--MethodOfValuation                 
## 27              ¦--PurposeOfGrant                    
## 28              °--Region

SetGraphStyle( nd, rankdir = "LR")
SetEdgeStyle( nd, arrowhead = "vee", color = "grey20", penwidth = 2 )
SetNodeStyle( nd, 
              style = "filled,rounded", 
              shape = "box", 
              fillcolor = "LightBlue", 
              fontname = "helvetica", 
              fontcolor="black",
              tooltip = GetDefaultTooltip )

# SetNodeStyle(acme$IT, fillcolor = "LightBlue", penwidth = "5px")
plot( nd )

# old version
find_group_names( table.name )

## [1] "//Form990ScheduleFPartII"  "//GrantsToOrgOutsideUSGrp"
## [3] "//GrantsToOrgsOutsideUS"

# new version
find_table_headers( table.name )

##  [1] "//Form990ScheduleFPartII/GrantsToOrgsOutsideUS"       
##  [2] "//Form990ScheduleFPartII/TotalNumberOf501c3Orgs"      
##  [3] "//Form990ScheduleFPartII/TotalNumberOfOtherOrgs"      
##  [4] "//Form990ScheduleFPartII/GrantRecordsMaintained"      
##  [5] "//GrantsToOrgOutsideUSGrp/RegionTxt"                  
##  [6] "//GrantsToOrgOutsideUSGrp/PurposeOfGrantTxt"          
##  [7] "//GrantsToOrgOutsideUSGrp/CashGrantAmt"               
##  [8] "//GrantsToOrgOutsideUSGrp/MannerOfCashDisbursementTxt"
##  [9] "//GrantsToOrgOutsideUSGrp/NonCashAssistanceAmt"       
## [10] "//GrantsToOrgOutsideUSGrp/DescriptionOfNonCashAsstTxt"
## [11] "//GrantsToOrgOutsideUSGrp/ValuationMethodUsedDesc"    
## [12] "//GrantsToOrgsOutsideUS/Form990ScheduleFPartII"       
## [13] "//GrantsToOrgsOutsideUS/GrantsToOrgsOutsideUS"

# manual
if( ! is.null(table.headers) )
{table.headers}

Build the Table

# create new folder
dir.create( folder.name )
setwd( folder.name )

# create sample
set.seed( 1234 )
sample.index <- sample_n( index, sample.size )
write.csv( index, "sample-index.csv", row.names=F )
sample.urls <- sample.index$URL

# erase existing log files
file.create("XPATH-LOG.txt") 
file.create("FAIL-LOG.txt")


start.build.time <- Sys.time()    # --------------------

results.list <- list()
xpath.list <- list()


for( i in 1:length( sample.urls ) )
{
  
  url <- sample.urls[i]
  
  results.list[[i]] <- 
    build_rdb_table_v2( url, 
                        table.name, 
                        table.headers=table.headers )
  
  xpath.list[[i]]   <- 
    get_table_xpaths( url, 
                      table.name, 
                      table.headers=table.headers )
  
  if( i %% 100 == 0 ){ print(i) }
  
}


end.build.time <- Sys.time()      # --------------------


df <- dplyr::bind_rows( results.list )


fail.log <- readLines( "FAIL-LOG.txt" )
all.xpaths <- xpath.list %>% unlist() %>% unique() %>% sort()
cat( all.xpaths, sep="\n", file="XPATH-LOG.txt" )

TABLE: SF-P02-T01-FRGN-ORG-GRANTS

## [1] "NUMBER OF RETURNS PROCESSED:   1000"
## [1] "TOTAL NUMBER OF TABLE ROWS:   5725"

TOTAL RUN TIME:

## Time difference of 30.28073 mins
## Time difference of -0.5046789 hours

Write to File

write.csv( df, paste0( folder.name, "/", table.name, ".csv" ), row.names=F )
head(df)

Are all names standardized?

names(df)

##  [1] "OBJECT_ID"                    "EIN"                         
##  [3] "NAME"                         "TAXYR"                       
##  [5] "FORMTYPE"                     "URL"                         
##  [7] "SF_02_FRGN_ORG_GRANT_REGION"  "SF_02_FRGN_ORG_GRANT_PURPOSE"
##  [9] "SF_02_FRGN_ORG_GRANT_DISBMT"  "SF_02_FRGN_ORG_GRANT_NC_MOV" 
## [11] "SF_02_FRGN_ORG_GRANT_AMT"     "SF_02_FRGN_ORG_GRANT_NC_AMT" 
## [13] "SF_02_FRGN_ORG_GRANT_NC_DESC"

Diagnostics

Missing or Extra XPATHS

Check for xpaths that were used in the tables but do not exist in the Concordance file.

c.table <- dplyr::filter( concordance, rdb_table == table.name ) 
xpath.concordance <- c.table$xpath
setdiff( all.xpaths, xpath.concordance )

## [1] ""

Capture to update the Concordance file.

these <- setdiff( all.xpaths, xpath.concordance )
cat( these, sep="\n", file=paste0( folder.name, "/MISSING-XPATHS.txt" ) )

Failes URLs

Check the status of failed URLs.

if( length(fail.log) > 0 )
{ 
  sapply( fail.log, get_message ) %>% 
  table() %>% 
  pander() 
} else {
  print("ALL URLs ARE VALID")
}

## [1] "ALL URLs ARE VALID"

XML Table Structure

Create Network Edgelists

The create_edgelist() functions v1 and v2 are in R/utils.R.

V1 retains the root nodes Return and ReturnData.

V2 drops the root nodes and uses form versions “990” and “990EZ” as the root to visualize the sets of tables separately.

# KEEP ROOT NOTES Return & ReturnData
el <- create_edgelist_v1( all.xpaths )

Edgelist format:

V1	V2
Return	ReturnData
ReturnData	IRS990
IRS990	Form990PartVIISectionAGrp
Form990PartVIISectionAGrp	PersonNm
Form990PartVIISectionAGrp	TitleTxt
Form990PartVIISectionAGrp	AverageHoursPerWeekRt
Form990PartVIISectionAGrp	IndividualTrusteeOrDirectorInd
Form990PartVIISectionAGrp	OfficerInd
Form990PartVIISectionAGrp	ReportableCompFromOrgAmt
Form990PartVIISectionAGrp	ReportableCompFromRltdOrgAmt

Visualize Tree Structure

# library( data.tree )
nd <- FromDataFrameNetwork( network=el )
print( nd )

##                                      levelName
## 1  Return                                     
## 2   °--ReturnData                             
## 3       °--IRS990ScheduleF                    
## 4           °--GrantsToOrgOutsideUSGrp        
## 5               ¦--CashGrantAmt               
## 6               ¦--DescriptionOfNonCashAsstTxt
## 7               ¦--MannerOfCashDisbursementTxt
## 8               ¦--NonCashAssistanceAmt       
## 9               ¦--PurposeOfGrantTxt          
## 10              ¦--RegionTxt                  
## 11              °--ValuationMethodUsedDesc

SetGraphStyle( nd, rankdir = "LR")
SetEdgeStyle( nd, arrowhead = "vee", color = "grey20", penwidth = 2 )
SetNodeStyle( nd, 
              style = "filled,rounded", 
              shape = "box", 
              fillcolor = "LightBlue", 
              fontname = "helvetica", 
              fontcolor="black",
              tooltip = GetDefaultTooltip )

# SetNodeStyle(acme$IT, fillcolor = "LightBlue", penwidth = "5px")
plot( nd )

# library( networkD3 )
# net <- ToDataFrameNetwork( nd, "name" )
# simpleNetwork( net, fontSize = 12 )
simpleNetwork( el, fontSize = 12 )

#plot with networkD3
radnet <- ToListExplicit( nd, unname = TRUE )
radialNetwork( radnet, fontSize = 12 )

XPATH Summary

Unique Xpaths for this RDB Table

t.xpaths <- 
  concordance %>%
  filter( rdb_table == table.name ) %>% 
  select( xpath ) %>% 
  unique() %>% 
  arrange( xpath ) %>% 
  mutate( id=row_number() ) 

t.xpaths %>% 
  kable( align="l" )

xpath	id
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/AmountOfCashGrant	1
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/AmountOfNonCashAssistance	2
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/DescriptionOfNonCashAssistance	3
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/MannerOfCashDisbursement	4
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/MethodOfValuation	5
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/PurposeOfGrant	6
/Return/ReturnData/IRS990ScheduleF/Form990ScheduleFPartII/GrantsToOrgsOutsideUS/Region	7
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/CashGrantAmt	8
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/DescriptionOfNonCashAsstTxt	9
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/MannerOfCashDisbursementTxt	10
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/NonCashAssistanceAmt	11
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/PurposeOfGrantTxt	12
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/RegionTxt	13
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgOutsideUSGrp/ValuationMethodUsedDesc	14
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/AmountOfCashGrant	15
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/AmountOfNonCashAssistance	16
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/DescriptionOfNonCashAssistance	17
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/MannerOfCashDisbursement	18
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/MethodOfValuation	19
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/PurposeOfGrant	20
/Return/ReturnData/IRS990ScheduleF/GrantsToOrgsOutsideUS/Region	21

Grouping Variables

Identify all unique groups used in xpaths.

if( is.null(table.headers) )
{ find_table_headers( table.name ) } else
table.headers

##  [1] "//Form990ScheduleFPartII/GrantsToOrgsOutsideUS"       
##  [2] "//Form990ScheduleFPartII/TotalNumberOf501c3Orgs"      
##  [3] "//Form990ScheduleFPartII/TotalNumberOfOtherOrgs"      
##  [4] "//Form990ScheduleFPartII/GrantRecordsMaintained"      
##  [5] "//GrantsToOrgOutsideUSGrp/RegionTxt"                  
##  [6] "//GrantsToOrgOutsideUSGrp/PurposeOfGrantTxt"          
##  [7] "//GrantsToOrgOutsideUSGrp/CashGrantAmt"               
##  [8] "//GrantsToOrgOutsideUSGrp/MannerOfCashDisbursementTxt"
##  [9] "//GrantsToOrgOutsideUSGrp/NonCashAssistanceAmt"       
## [10] "//GrantsToOrgOutsideUSGrp/DescriptionOfNonCashAsstTxt"
## [11] "//GrantsToOrgOutsideUSGrp/ValuationMethodUsedDesc"    
## [12] "//GrantsToOrgsOutsideUS/Form990ScheduleFPartII"       
## [13] "//GrantsToOrgsOutsideUS/GrantsToOrgsOutsideUS"

Groups can then be used to extract the 1:M RDB tables:

if( is.null(table.headers) )
{ table.headers <- find_table_headers( table.name ) }
demo.url <- "https://nccs-efile.s3.us-east-1.amazonaws.com/xml/201913079349300206_public.xml"
doc <- xml2::read_xml( demo.url )
xml2::xml_ns_strip( doc )
df.demo <- get_table_v2( doc, table.name, table.headers )
df.demo

## NULL

Standardize Variable Names

Mapping of xml variable names to concordance variable names:

get_var_map( table.name=table.name )

A table with proper names:

v.map <- get_var_map( table.name=table.name )

if( ! is.null(df.demo) )
{re_name( df.demo, v.map ) %>% as.data.frame()}

Print Table Structure

Demo of functions - demo data only.

[[1]]
<Form990PartVIISectionAGrp>
  <PersonNm>
    {text}
  <TitleTxt>
    {text}
  <AverageHoursPerWeekRt>
    {text}
  <IndividualTrusteeOrDirectorInd>
    {text}
  <OfficerInd>
    {text}
  <ReportableCompFromOrgAmt>
    {text}
  <ReportableCompFromRltdOrgAmt>
    {text}
  <OtherCompensationAmt>
    {text}

node.list[[1]] %>% xml_parent() %>% xmltools::xml_view_trees()

 (1) ------------

 +-- Form990PartVIISectionAGrp
  +-- PersonNm
  +-- TitleTxt
  +-- AverageHoursPerWeekRt
  +-- IndividualTrusteeOrDirectorInd
  +-- OfficerInd
  +-- ReportableCompFromOrgAmt
  +-- ReportableCompFromRltdOrgAmt
  ++-- OtherCompensationAmt


 (2) ------------

 +-- Form990PartVIISectionAGrp
  +-- PersonNm
  +-- TitleTxt
  +-- AverageHoursPerWeekRt
  +-- IndividualTrusteeOrDirectorInd
  +-- OfficerInd
  +-- ReportableCompFromOrgAmt
  +-- ReportableCompFromRltdOrgAmt
  ++-- OtherCompensationAmt

TABLE NAME: SF-P02-T01-FRGN-ORG-GRANTS

Last compiled on 04 April, 2023