Creating Network

1. Check if valid email
2. Check if "from" is valid name and extract it
3. check if the name is in the network by reversing the first and last name. flip if exists
4. For each to and cc
     1. Check if "to" is valid name and extract it
     2. Check if the name is in the network by reversing the first and last name. flip if exists
     3. if all checks were good, insert new edge "from" -> "to" into network

Getting rid of "spam" Emails

1. blacklist
2. If its cc'd to more than x people, don't use it in the network

Extract Name

1. remove troubling characters such as: & . ? * ( ) '
2. if digits, or @, then its not a name
3. remove titles such as: Mr Mrs Dr ans suffixes such as: Jr Sr
4. if more than 3 words, then its not a name
5. Convert all names to same format: (note case)
Valid Formats: (size is 2 or 3)
     a. LAST, FIRST
     b. FIRST LAST
     c. LAST, FIRST MIDDLE|INITIAL.
     d. FIRST MIDDLE|INITIAL.

Analysis

10-24-00, 11-29-00, and 12-20-00 (Skilling named as CEO) Network

Date	Blacklist	Mass Email Removal at	# Nodes	#Edges
10/24/00	Yes	10	436	640
10/24/00	No	10	441	1090
10/24/00	Yes	20	541	1379
11/29/00	Yes	5	397	451
11/29/00	Yes	10	567	728
11/29/00	No	10	572	733
11/29/00	Yes	20	770	1154
11/29/00	No	20	775	1158
12/20/00	Yes	10	2080	6016
12/20/00	No	10	2099	6043
12/20/00	No	20	2675	8719

Note: Using a blacklist can delete more than just those "names", since people may only converse with the blacklist.

Name Problems:
1. There are nodes that should have been, but were not listed in the blacklist
2. Some names are more than 3 words. Example: Janet De La Paz
In 12-20-00 (blacklist and x=10) Network, 26/2080 had errors; a mere 1.2% - Can be reduced even further by adding more names to the blacklist.

Top 10 - 12-20-00 (blacklist and x=10)

#	person	degree
1	Jeff Dasovich	145
2	Vince Kaminski	121
3	Sara Shackleton	120
4	Steven Kean	119
5	Tana Jones	112
6	Kay Mann	93
7	Jeffrey Shankman	90
8	Mark Taylor	86
9	Chris Germany	85
10	David Delainey	83
...
35	Jeff Skilling	43

Top 10 - 11-29-00 (blacklist and x=10)

#	person	degree
1	Tana Jones	32
2	Sara Shackleton	27
3	Jeff Dasovich	24
4	Vince Kaminski	23
5	Susan Scott	20
6	Chris Germany	19
7	Kate Symes	19
8	Mike Mcconnell	15
9	Jeffrey Shankman	13
10	Karen Denne	13
...
133	Jeff Skilling	3

Networks

11-29-00 Network (x=10)
12-20-00 Network (x=10)
11-29-00 Strongly Connected Component Subnet