I have saved all posts in the newsgroup rec.sport.unicycling since late 2000, and noticed some trends. One of my pastimes is trivial or pointless data analysis - the wording and definition of this hobby do vary. I decided to combine the two, and here are some results. Hey guys and girls, your posting habits are revealed without mercy...

Disclaimers and how-I-did-it are at the bottom of the page.

In 2001, a total of 7709 messages were posted to the newsgroup, by a total of 638 posters. The overall number of posts per day has varied quite a bit. There was only one sad day with zero posts (25 January), while the record day was 14 November, sporting 107 posts. As a trend, the newsgroup has clearly become busier over the year. In the picture below, the number of posts has been smoothed using a running average (using a one-week window so no effect of weekly trends).

From above picture someone thought to see spikes at the beginning and end of each month. The picture below shows that this is not really the case. The magnitude of the variations is what you would expect statistically. Throw the paycheck theory out of the window. (Oops! Not the paycheck!)

It is noteworthy that on Saturday and Sunday posting almost comes to a grinding halt. I guess people are too busy unicycling in the weekend to post (or too tired from riding to post in the evening). Other than that, posts are pretty evenly distributed.

The table below gives the 50 most prolific posters (in total, 638 people have posted at least once in 2001). The order has been based firstly on the number of posts during the year, secondly (if that was inconclusive) on the average number of lines per post, and thirdly on the (low) number of lines quoted, i.e. more original content.

The meaning of the other columns is as follows, taking John Foss as an example: John posted 455 times in 2001, the text body of his posts consisted of 32 lines on average, of which 18.8% was quoted from previous messages. 5.90% of all the posts in the group were from John, with which he occupied 7.30% of the total number of bytes posted in 2001. Finally, 9.91% of the newsgroup posts in January were posted by John, 7.29% in February, etc.

#

name

# of msgs

avg msg lines

%quoted

%group msgs

%group bytes

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

1

John Foss

455

32

18.80%

5.90%

7.30%

9.91

7.29

5.66

5.37

6.15

7.63

6.06

7.95

7.16

5.96

5.31

2.89

2

Jeff Lutkus

275

27

29.90%

3.57%

4.20%

7.21

2.08

1.37

2.93

2.74

7.2

4.19

5.43

2.71

1.95

2.09

4.73

3

Klaas Bil

270

27

37.60%

3.50%

3.20%

0.9

2.08

4.49

2.44

2.39

3.11

3.57

2.71

3.29

3.9

3.54

5.17

4

Kris Holm

223

40

42.20%

2.89%

4.10%

2.25

2.78

2.54

1.22

3.93

4.66

5.59

2.33

2.51

2.71

3.62

0.44

5

David Stone

213

30

32.30%

2.76%

3.20%

0

0.69

4.3

3.66

3.93

5.08

4.35

2.71

0.97

1.3

3.22

1.4

6

Nathan Hoover

188

32

41.60%

2.44%

2.80%

2.7

2.78

2.93

3.9

2.91

3.95

3.26

2.33

2.51

2.6

1.69

0.61

7

Jack Halpern

185

42

0.50%

2.40%

3.50%

2.7

1.39

2.73

4.63

1.03

3.95

6.06

0.39

1.55

0.65

3.3

1.05

8

Christopher Grove

182

28

23.90%

2.36%

2.10%

0

0

0

0

0

0

0

0

0

5.74

10.06

0.35

9

John Childs

171

36

35.90%

2.22%

2.90%

4.05

3.82

1.37

2.93

2.56

1.55

1.25

1.36

1.55

4.01

1.93

1.93

10

rhysling

157

22

5.40%

2.04%

1.20%

0

0

0

0

0

0

0

0

0

1.73

6.68

5.08

11

Chris Reeder

142

26

34.80%

1.84%

1.60%

3.6

4.17

2.93

2.2

2.91

3.11

3.26

1.74

2.32

0

0.32

1.14

12

Sarah Miller

138

25

13.80%

1.79%

1.70%

2.25

3.47

2.54

2.44

1.71

1.84

2.48

1.74

3.09

2.49

0.72

0.35

13

Neil Dunlop

137

44

27.80%

1.78%

2.30%

0.45

3.13

1.76

0.98

2.22

1.84

3.57

2.33

3.29

1.08

1.05

1.14

14

harper

137

18

6.90%

1.78%

1.00%

0

0

0

0

0

0

0

0

0

3.14

2.66

6.57

15

Joe Marshall

118

35

27.30%

1.53%

2.00%

0.45

3.47

1.17

0.98

1.54

0.42

0.16

1.36

1.36

2.82

3.06

0.53

16

Danny Colyer

102

28

16.90%

1.32%

1.20%

1.35

0.35

1.56

3.17

3.59

1.55

1.24

1.55

1.16

0.76

0.89

0.44

17

Olivier Paugam

95

39

7.90%

1.23%

1.40%

0

0

0.59

0.73

1.71

2.12

2.02

2.33

2.71

1.3

0.97

0.09

18

Nycjoe@aol.com

91

37

23.50%

1.18%

1.80%

0

0.35

2.73

2.2

0.85

1.98

2.64

0.58

0.97

0.22

0.97

0.79

19

Don Tai

89

43

41.70%

1.15%

1.90%

0

0

0.98

0

0.85

0.71

1.55

3.49

2.71

1.95

1.13

0

20

Roger

86

48

42.80%

1.12%

1.70%

1.35

2.78

2.34

2.44

1.71

0.99

0.93

1.94

2.9

0.33

0.16

0

21

Graham W. Boyes - TOAO.net

86

27

26.80%

1.12%

0.80%

0

0.35

0

0.73

2.56

2.97

2.95

1.16

0.39

0.43

0.89

0.44

22

jagur

85

11

9.80%

1.10%

0.20%

0

0

0

0

0

0

0

0

0

0

0.97

6.39

23

John Hooten

83

42

32.60%

1.08%

1.80%

2.7

2.08

1.76

2.2

1.03

0.28

0.78

2.33

2.71

0.76

0.16

0.44

24

Bruce & Mary Edwards

75

31

45.10%

0.97%

1.10%

0

0

0

0

0

1.27

1.55

1.36

2.51

1.19

1.21

0.88

25

Maxfield D

73

12

2.80%

0.95%

0.40%

1.8

0.35

2.15

0.73

0.17

0.28

0.31

0.78

0.58

1.3

0.89

1.66

26

uni-man-dan

70

16

7.20%

0.91%

0.30%

0

0

0

0

0

0

0

0

0

0.22

2.01

3.77

27

dingeman

70

16

21.10%

0.91%

0.50%

0

0

3.13

1.22

2.05

1.69

0.78

0.58

1.55

0.65

0.16

0.09

28

Gilby

68

18

8.10%

0.88%

0.50%

0

0

0

2.44

1.71

0.14

1.55

1.36

0.39

0.65

0.89

0.96

29

Greg House

67

22

32.10%

0.87%

0.60%

0.45

0

1.76

2.93

2.22

1.55

2.33

0.58

0.39

0.11

0

0

30

Trevor Coultart

66

20

25.50%

0.86%

0.40%

0.9

0.35

2.15

0.24

1.03

3.11

0.93

0.19

0.19

0.33

0.56

0.44

31

teachndad

64

24

5.20%

0.83%

0.50%

0

0

0

0

0

0

0.31

0.97

2.71

1.63

0.72

1.66

32

Andrew Feldhaus

62

24

17.20%

0.80%

0.60%

0

0

0

0

1.54

0

0

1.16

1.93

1.52

0.48

1.49

33

Mark Stephens

58

28

43.40%

0.75%

0.80%

1.35

2.08

1.56

2.44

1.03

0.99

0.78

0.58

0.58

0

0.16

0.44

34

Peter Bier

57

31

1.30%

0.74%

0.80%

2.7

2.08

2.54

1.22

1.54

0.99

0.31

0.39

0.77

0.22

0

0.09

35

Michael Grant

56

32

38.50%

0.73%

0.70%

0

0

0

0

0

0

0

0

0

2.38

1.45

1.4

36

Nicholas Price

55

37

39.50%

0.71%

0.90%

0

0

0.2

0.49

0.85

0.71

1.4

0.97

0.77

0.98

0.64

0.61

37

Trev

51

24

12.60%

0.66%

0.60%

0

2.08

4.88

0.49

0.68

0.42

0

0

0

0.98

0.08

0.09

38

Kevin Gilbertson

51

21

21.90%

0.66%

0.50%

0.9

1.04

0.78

0.49

0.68

0.28

0.62

0.19

1.74

0.76

0.32

0.79

39

John

50

27

38.60%

0.65%

0.50%

0

0

0

0

0

0

0

0

1.93

0.87

0.89

1.84

40

AccordNSX

48

12

8.60%

0.62%

0.20%

0

0

0

0

0

0

0

0

0

0

0

4.2

41

skunker

47

18

6.10%

0.61%

0.30%

0

0

0

0

0

0

0

0

0

1.41

1.53

1.31

42

Arnold the Aardvark

47

17

6.80%

0.61%

0.30%

0

0.69

0.2

0

0.34

0.28

0

0.39

0.58

2.38

0.64

0.44

43

Wayne van Wijk

45

26

17.30%

0.58%

0.40%

1.8

0.35

0.2

0.48

1.37

0.85

0.63

0.39

1.54

0.65

0.08

0.44

44

Carl Hoyer

42

64

25.80%

0.54%

1.20%

1.35

1.04

0.2

0

0.85

1.41

0.47

1.36

0.58

0

0.48

0.09

45

jeff d tuttle

42

60

48.80%

0.54%

1.10%

0

0

0.2

0.49

0.17

0.28

0.16

0.19

0.77

1.3

0.81

0.7

46

David Straitjacket

41

44

43.50%

0.53%

0.70%

0

0

0

0

0

0

0

0

2.13

1.73

0.72

0.44

47

Roger Davies

39

59

51.10%

0.51%

1.00%

0.45

0.35

0.78

1.22

0.51

0.14

0

0.19

0.19

0.54

1.05

0.35

48

Christian St-Pierre

39

34

40.10%

0.51%

0.60%

0

0

0

0

0

0

0.31

0

0.97

1.19

1.21

0.53

49

Import Car Fan

39

30

25.30%

0.51%

0.50%

0

0

0

0

0

0

0.31

0.58

1.16

0.54

0.32

1.66

50

Dylan Wallinger

36

9

5.30%

0.47%

0.20%

0

0

0

0

0

0

0

0

0

0.98

0.81

1.49

Eight out of the top-10 posters were active throughout the year. In October, two Christophers came out of the blue "with a dot" who almost took over the newsgroup :-) Here's the activity log of the ten people that between themselves posted about 29% of all the 2001 posts:

The posts in 2001 were devoted to a total of 2030 threads, or rather subjects since a post that hasn't been replied to might technically not be called a thread. Regulars will recognise most if not all of the discussions. A thread on our ever-hot topic of comments from non-riders comes out as number one. But the new trials competition format proposed by Kris Holm would firmly come on top if the order were based on number of bytes - aaargh thoroughly solid-worded arguments were flung onto the electronic highway. Anyway, here's the top 10:

subject

number of posts

% of group posts

% of group bytes

1

We are not GEEKS! or WEIRDOS

55

0.71%

0.69%

2

unicycle.com

43

0.56%

0.31%

3

Doin' it on a uni??????

41

0.53%

0.29%

4

Intelligent Unicyclists

35

0.45%

0.55%

5

Soon-to-be new learner wants advice on expectations

35

0.45%

0.54%

6

Unicycle synonyms

34

0.44%

0.53%

7

[rsu] Too Many Unicycles?

30

0.39%

0.38%

8

New Trials Unicycle Competition Format- IMPORTANT!!!

29

0.38%

1.37%

9

pedals

29

0.38%

0.19%

10

What got you here

28

0.36%

0.60%

Disclaimers and how-I-did-it

I have relied on the news server of my ISP to view the newsgroup - but the server may have missed some posts.
I have used Free Agent 1.21/32 to view and save all messages, excluding spam but including replies of genuine posters to spam (Don Tai did an interesting experiment). Some spam may have been saved and counted in, while some genuine posts may have been missed, particularly because in my news reader threads are kept together and ordered by date of original post, so when people respond to a very old post it's way up and I may miss it. Yes I think I could change that - I just haven't.
Occasionally, messages show up twice or more. (People reading via the forum seem to experience that more often.) In that case messages are double (triple etc.) counted.
The date assigned to a post is not trivial. AFAIK, each post is date/time-stamped by the poster's own computer at the time of sending. Free Agent does convert that date/time stamp to my time zone (Central European). I don't know how Day Light Saving time is accounted for. We do have DST here during the summer, but some posters may not.
I have generated raw statistics using QDNStatsV2 (RtS#2) by Brett O'Callaghan. QDNStats pretends to generate web-ready reports but I think they're still quite tedious. Other than that it works flawlessly as far as I can see.
QDNStats distinguishes posters by their names. Not so strange, most people do the same but QDNStats is very picky. For instance, "Graham W. Boyes - TOAO.net", "G.W. Boyes - TOAO.net", "Graham W.Boyes" and "Graham W. Boyes" are four different posters. See how subtle it gets? I noticed at least six people whom I suspect posted under different names. I made no effort to correct this, partly because it is not easy, and also how can I be sure that it indeed involves a single person? I apologise, particularly to the one individual who answered my request to notify me of different posting names.
I used Excel to post-process the QDNStats results, and Paint Shop Pro to polish the graphics. Finally, I just used Word to generate my web pages...

Back to Unicycling statistics.