diff --git a/doc/figs/ghtorrent-process.pdf b/doc/figs/ghtorrent-process.pdf
deleted file mode 100644
index 3eafb44..0000000
Binary files a/doc/figs/ghtorrent-process.pdf and /dev/null differ
diff --git a/doc/figs/ghtorrent-process.pic b/doc/figs/ghtorrent-process.pic
deleted file mode 100644
index 7ee72a9..0000000
--- a/doc/figs/ghtorrent-process.pic
+++ /dev/null
@@ -1,48 +0,0 @@
-.PS
-copy "sequence.pic";
-
-# Object definition
-object(E,"EventMirror");
-object(G,"GHTorrent");
-object(R,"Retriever");
-object(P,"Persister");
-object(A,"APIClient");
-object(GH,"GithubAPI");
-step();
-
-# Message exchange
-active(E);
-message(E,G,"ensure_pullreq(p, id)");
-active(G);
-message(G,G,"ensure_project_owner(p)");
-message(G,G,"ensure_project(p)");
-message(G,R,"retrieve_pullreq(p, id)");
-active(R);
-message(R,P,"find_pullreq(p, id)");
-active(P);
-message(R,A,"get_pullreq(p, id)");
-active(A);
-message(A,GH, "GET /repo/pullreq/id");
-active(GH);
-return_message(GH,A)
-inactive(GH);
-message(A,A,"cache_http()");
-return_message(A,R)
-inactive(A);
-message(R,P,"store_pullreq(p, id)");
-return_message(P,R);
-inactive(P);
-return_message(R,G, "pull_req_json");
-inactive(R);
-message(G,G,"proc_save_pullreq(p, id)");
-return_message(G,E, "pull_req");
-
-complete(E);
-complete(G);
-complete(R);
-complete(P);
-complete(A);
-complete(GH);
-
-.PE
-
diff --git a/doc/figs/github-growth.pdf b/doc/figs/github-growth.pdf
deleted file mode 100644
index 12f5032..0000000
Binary files a/doc/figs/github-growth.pdf and /dev/null differ
diff --git a/doc/figs/sequence.pic b/doc/figs/sequence.pic
deleted file mode 100644
index 35e443e..0000000
--- a/doc/figs/sequence.pic
+++ /dev/null
@@ -1,407 +0,0 @@
-#/usr/bin/pic2plot -Tps
-#
-# Pic macros for drawing UML sequence diagrams
-#
-# (C) Copyright 2004-2005 Diomidis Spinellis.
-#
-# Permission to use, copy, and distribute this software and its
-# documentation for any purpose and without fee is hereby granted,
-# provided that the above copyright notice appear in all copies and that
-# both that copyright notice and this permission notice appear in
-# supporting documentation.
-#
-# THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
-# WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
-# MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
-#
-# $Id$
-#
-
-
-# Default parameters (can be redefined)
-
-# Spacing between messages
-spacing = 0.25;
-# Active box width
-awid = .1;
-# Box height
-boxht = 0.3;
-# Commend folding
-corner_fold=awid
-# Comment distance
-define comment_default_move {up 0.25 right 0.25};
-# Comment height
-comment_default_ht=0.5;
-# Comment width
-comment_default_wid=1;
-# Underline object name
-underline=1;
-
-# Create a new object(name,label)
-define object {
-	$1: box $2; move;
-	# Could also underline text with \mk\ul\ul\ul...\rt
-	if (underline) then {
-		line from $1.w + (.1, -.07) to $1.e + (-.1, -.07);
-	}
-	move to $1.e;
-	move right;
-	# Active is the level of activations of the object
-	# 0 : inactive : draw thin line swimlane
-	# 1 : active : draw thick swimlane
-	# > 1: nested : draw nested swimlane
-	active_$1 = 0;
-	lifestart_$1 = $1.s.y;
-}
-
-# Create a new external actor(name,label)
-define actor {
-	$1: [
-		XSEQC: circle rad 0.06;
-		XSEQL: line from XSEQC.s down .12;
-		line from XSEQL.start - (.15,.02) to XSEQL.start + (.15,-.02);
-		XSEQL1: line from XSEQL.end left .08 down .15;
-		XSEQL2: line from XSEQL.end right .08 down .15;
-		line at XSEQC.n invis "" "" "" $2;
-	]
-	move to $1.e;
-	move right;
-	active_$1 = 0;
-	lifestart_$1 = $1.s.y - .05;
-}
-
-# Create a new placeholder object(name)
-define placeholder_object {
-	$1: box invisible;
-	move;
-	move to $1.e;
-	move right;
-	active_$1 = 0;
-	lifestart_$1 = $1.s.y;
-}
-
-define pobject {
-	placeholder_object($1);
-}
-
-define extend_lifeline {
-	if (active_$1 > 0) then {
-                # draw the left edges of the boxes
-		move to ($1.x - awid/2, Here.y);
-		for level = 1 to active_$1 do {
-			line from (Here.x, lifestart_$1) to Here;
-			move right awid/2
-		}
-
-                # draw the right edge of the innermost box
-		move right awid/2;
-		line from (Here.x, lifestart_$1) to Here;
-	} else {
-		line from ($1.x, lifestart_$1) to ($1.x, Here.y) dashed;
-	}
-	lifestart_$1 = Here.y;
-}
-
-# complete(name)
-# Complete the lifeline of the object with the given name
-define complete {
-	extend_lifeline($1)
-	if (active_$1) then {
-		# draw bottom of all active boxes
-		line right ((active_$1 + 1) * awid/2) from ($1.x - awid/2, Here.y);
-	}
-}
-
-# Draw a message(from_object,to_object,label)
-define message {
-	down;
-	move spacing;
-	# Adjust so that lines and arrows do not fall into the
-	# active box.  Should be .5, but the arrow heads tend to
-	# overshoot.
-	if ($1.x <= $2.x) then {
-		off_from = awid * .6;
-		off_to = -awid * .6;
-	} else {
-		off_from = -awid * .6;
-		off_to = awid * .6;
-	}
-
-        # add half a box width for each level of nesting
-        if (active_$1 > 1) then {
-                off_from = off_from + (active_$1 - 1) * awid/2;
-        }
-
-        # add half a box width for each level of nesting
-        if (active_$2 > 1) then {
-                off_to = off_to + (active_$2 - 1) * awid/2;
-        }
-
-	if ($1.x == $2.x) then {
-		arrow from ($1.x + off_from, Here.y) right then down .25 then left $3 ljust " " " " " " ;
-	} else {
-		arrow from ($1.x + off_from, Here.y) to ($2.x + off_to, Here.y) $3 " ";
-	}
-}
-
-# Display a lifeline constraint(object,label)
-define lifeline_constraint {
-        off_from = awid;
-        # add half a box width for each level of nesting
-        if (active_$1 > 1) then {
-                off_from = off_from + (active_$1 - 1) * awid/2;
-        }
-
-	box at ($1.x + off_from, Here.y) invis $2 ljust " " ;
-}
-
-define lconstraint {
-	lifeline_constraint($1,$2);
-}
-
-# Display an object constraint(label)
-# for the last object drawn
-define object_constraint {
-	{ box invis with .s at last box .nw $1 ljust; }
-}
-
-define oconstraint {
-	object_constraint($1);
-}
-
-# Draw a creation message(from_object,to_object,object_label)
-define create_message {
-	down;
-	move spacing;
-	if ($1.x <= $2.x) then {
-		off_from = awid * .6;
-		off_to = -boxwid * .51;
-	} else {
-		off_from = -awid * .6;
-		off_to = boxwid * .51;
-	}
-
-        # add half a box width for each level of nesting
-        if (active_$1 > 1) then {
-                off_from = off_from + (active_$1 - 1) * awid/2;
-        }
-
-	# See comment in destroy_message
-	XSEQA: arrow from ($1.x + off_from, Here.y) to ($2.x + off_to, Here.y) "«create»" " ";
-	if ($1.x <= $2.x) then {
-		{ XSEQB: box $3 with .w at XSEQA.end; }
-	} else {
-		{ XSEQB: box $3 with .e at XSEQA.end; }
-	}
-	{
-		line from XSEQB.w + (.1, -.07) to XSEQB.e + (-.1, -.07);
-	}
-	lifestart_$2 = XSEQB.s.y;
-	move (spacing + boxht) / 2;
-}
-
-define cmessage {
-	create_message($1,$2,$3);
-}
-
-# Draw an X for a given object
-define drawx {
-	{
-	line from($1.x - awid, lifestart_$1 - awid) to ($1.x + awid, lifestart_$1 + awid);
-	line from($1.x - awid, lifestart_$1 + awid) to ($1.x + awid, lifestart_$1 - awid);
-	}
-}
-
-# Draw a destroy message(from_object,to_object)
-define destroy_message {
-	down;
-	move spacing;
-	# The troff code is \(Fo \(Fc
-	# The groff code is also \[Fo] \[Fc]
-	# The pic2plot code is \Fo \Fc
-	# See http://www.delorie.com/gnu/docs/plotutils/plotutils_71.html
-	# To stay compatible with all we have to hardcode the characters
-	message($1,$2,"«destroy»");
-	complete($2);
-	drawx($2);
-}
-
-define dmessage {
-	destroy_message($1,$2);
-}
-
-# An object deletes itself: delete(object)
-define delete {
-	complete($1);
-	lifestart_$1 = lifestart_$1 - awid;
-	drawx($1);
-}
-
-# Draw a message return(from_object,to_object,label)
-define return_message {
-	down;
-	move spacing;
-	# See comment in message
-	if ($1.x <= $2.x) then {
-		off_from = awid * .6;
-		off_to = -awid * .6;
-	} else {
-		off_from = -awid * .6;
-		off_to = awid * .6;
-	}
-
-        # add half a box width for each level of nesting
-        if (active_$1 > 1) then {
-                off_from = off_from + (active_$1 - 1) * awid/2;
-        }
-
-        # add half a box width for each level of nesting
-        if (active_$2 > 1) then {
-                off_to = off_to + (active_$2 - 1) * awid/2;
-        }
-
-	arrow from  ($1.x + off_from, Here.y) to ($2.x + off_to, Here.y) dashed $3 " ";
-}
-
-define rmessage {
-	return_message($1,$2,$3);
-}
-
-# Object becomes active
-# Can be nested to show recursion
-define active {
-	extend_lifeline($1);
-	# draw top of new active box
-	line right awid from ($1.x + (active_$1 - 1) * awid/2, Here.y);
-	active_$1 = active_$1 + 1;
-}
-
-# Object becomes inactive
-# Can be nested to show recursion
-define inactive {
-	extend_lifeline($1);
-	active_$1 = active_$1 - 1;
-	# draw bottom of innermost active box
-	line right awid from ($1.x + (active_$1 - 1) * awid/2, Here.y);
-}
-
-# Time step
-# Useful at the beginning and the end
-# to show object states
-define step {
-	down;
-	move spacing;
-}
-
-# Switch to asynchronous messages
-define async {
-	arrowhead = 0;
-	arrowwid = arrowwid * 2;
-}
-
-# Switch to synchronous messages
-define sync {
-	arrowhead = 1;
-	arrowwid = arrowwid / 2;
-}
-
-# same as lifeline_constraint, but Text and empty string are exchanged.
-define lconstraint_below{
-        off_from = awid;
-        # add half a box width for each level of nesting
-        if (active_$1 > 1) then {
-                off_from = off_from + (active_$1 - 1) * awid/2;
-        }
-
-	box at ($1.x + off_from, Here.y) invis "" $2 ljust;
-}
-
-# begin_frame(left_object,name,label_text);
-define begin_frame {
-	# The lifeline will be cut here
-	extend_lifeline($1);
-	# draw the frame-label
-	$2: box $3 invis with .n at ($1.x, Here.y);
-	d = $2.e.y - $2.se.y;
-	line from $2.ne to $2.e then down d left d then to $2.sw;
-	# continue the lifeline below the frame-label
-	move to $2.s;
-	lifestart_$1 = Here.y;
-}
-
-# end_frame(right_object,name);
-define end_frame {
-	# dummy-box for the lower right corner:
-	box invis "" with .s at ($1.x, Here.y);
-	# draw the frame
-	frame_wid = last box.se.x - $2.nw.x
-	frame_ht = - last box.se.y + $2.nw.y
-	box with .nw at $2.nw wid frame_wid ht frame_ht;
-	# restore Here.y
-	move to last box.s;
-}
-
-# comment(object,[name],[line_movement], [box_size] text);
-define comment {
-	old_y = Here.y
-	# draw the first connecting line, at which's end the box wil be positioned
-	move to ($1.x, Here.y)
-	if "$3" == "" then {
-		line comment_default_move() dashed;
-	} else {
-		line $3 dashed;
-	}
-
-	# draw the box, use comment_default_xx if no explicit
-	# size is given together with the text in parameter 4
-	old_boxht=boxht;
-	old_boxwid=boxwid;
-	boxht=comment_default_ht;
-	boxwid=comment_default_wid;
-	if "$2" == "" then {
-		box invis $4;
-	} else {
-		$2: box invis $4;
-	}
-	boxht=old_boxht;
-	boxwid=old_boxwid;
-
-	# draw the frame of the comment
-	line from       last box.nw \
-		to          last box.ne - (corner_fold, 0) \
-		then to last box.ne - (0, corner_fold) \
-		then to last box.se \
-		then to last box.sw \
-		then to last box.nw ;
-	line from       last box.ne - (corner_fold, 0) \
-		to          last box.ne - (corner_fold, corner_fold) \
-		then to last box.ne - (0, corner_fold) ;
-
-	# restore Here.y
-	move to ($1.x, old_y)
-}
-
-# connect_to_comment(object,name);
-define connect_to_comment {
-	old_y = Here.y
-	# start at the object
-	move to ($1.x, Here.y)
-	# find the best connection-point of the comment to use as line-end
-	if $1.x < $2.w.x then {
-		line to $2.w dashed;
-	} else {
-		if $1.x > $2.e.x then {
-			line to $2.e dashed;
-		} else {
-			if Here.y < $2.s.y then {
-				line to $2.s dashed;
-			} else {
-				if Here.y > $2.n.y then {
-					line to $2.n dashed;
-				}
-			}
-		}
-	}
-	# restore Here.y
-	move to ($1.x, old_y)
-}
diff --git a/doc/latex/ghtorrent-data.tex b/doc/latex/ghtorrent-data.tex
index 5171466..ab857e6 100644
--- a/doc/latex/ghtorrent-data.tex
+++ b/doc/latex/ghtorrent-data.tex
@@ -14,7 +14,7 @@
 Software Engineering Research Group\\
 Delft University of Technology\\
 Delft, The Netherlands\\
-Email: g.gousios@tudelft.nl}
+g.gousios@tudelft.nl}
 }
 
 \maketitle
@@ -175,7 +175,7 @@ \section{Data Collection}
 This design choice has been very important since it makes our data stores append
 only and the results of each step memoizable and therefore cachable.
 
-\subsection*{Data schema}
+\subsection{Data schema}
 
 The data schema can be seen in Figure~\ref{fig:entities}. Following Github's
 {\sc api} organization, most entities belong to a {\sf project} and contain
@@ -190,7 +190,7 @@ \subsection*{Data schema}
 field contains the unique identifier to the raw entity in the Mongo{\sc
 db} database.
 
-\subsection*{Workers for Data}
+\subsection{Workers for Data}
 
 The data collection was designed from the beginning as a decentralized process.
 Decentralization is mediated using the worker queue model; a message producer