bots = [ n for n in pages_editors_graph.nodes() if ("bot" in n.lower()) and not("p:") in n ]
print len(bots)

data = []

bots = sorted(bots, key=lambda b: -pages_editors_graph.node[b]["revisions"])

col_names = ["bot name", "revisions", "edited pages", "revisions/page", "mean contribution", "first activity", "last activity"]

g = pages_editors_graph

for b in bots:
    link = u"<a href=\"http://en.wikipedia.org/wiki/User:{0}\" target=\"_blank\">{0}</a>".format(b.split(":")[1])
    r  = g.node[b]["revisions"]
    p  = len(pages_editors_graph[b])
    m  = np.mean(map(lambda p: float(g[b][p]["revisions"])/float(g.node[p]["revisions"]), g[b]))
    rp = float(r)/float(p)
    f_rev = g.node[b]["first revision"]
    l_rev = g.node[b]["last revision"]
    data.append([link, r, p, rp, m, f_rev, l_rev])

count_all_rev = sum([ g.node[p]["revisions"] for p in pages_editors_graph.nodes() if g.node[p]["type"] == "page" ])
count_bots_rev = sum([ b[1] for b in data ])

print "total number of revisions: %s" % (count_all_rev)
print "revisions by all bots: %s (%s)" % (count_bots_rev, float(count_bots_rev)/float(count_all_rev) * 100)

display(table_to_html(data, col_names))

df = pd.DataFrame(data=data, index=[ b.split(":")[1] for b in bots ], columns=col_names)

df.drop("bot name",1, inplace=True)

df.head()

df.plot("edited pages", "revisions", kind="scatter");