Various Statistics for 2008 • VideoSift: Online Video *Quality Control

It Puts the Blog on the Sift or Else It Gets the Hose Again

Various Statistics for 2008

posted by lucky760

10 months 2 weeks ago • 884 views

On the subject of statistics... Looking at Dag's "Visualizing 2000 Visitors an Hour" blog post I thought it would be interesting to throw together some graphs for various stats for the entire year of 2008. They are presented below in separate graphs keeping categories with similar values together for scale.

Included are: videos published, video down votes, video up votes, comments posted, comment up votes, comment down votes, comment spam votes, new users, banned users, Sift Talks posts, and blogs posted.

Note the spike in April for new users is due to a spammer who created a few hundred sock puppet accounts

comments (10)

bookmarks (0)

Stumble This

Comments

Very interesting! Thanks!

Seems that almost every graph is climbing until November.
Might be due to the rising amount of election related clips that were posted towards that date?

Or can you see another reason for that in your stats?

written by maatc

| 10 months 2 weeks ago | CH

▲ 2 ▼ | flag spam (0)

looks to me like you've got a database chain tensioner failing on you there around October, you might need to replace a drive gear and get a set of new rails and make sure you guys check that MySQL has enough oil always

written by deathcow

| 10 months 2 weeks ago | CH

▲ 7 ▼ | flag spam (0)

Is it weird that I spend a lot of time customizing my excel charts so they don't look like excel?

written by 14087 | 10 months 2 weeks ago | CH

▲ 2 ▼ | flag spam (0)

Wow, proof that behavior nazis and all these cliquey comment downvoters are actually harmful. You'll notice all the comment abuse by marking spam and downvoting leads to a drop off of activity.

I mean a few active users leaving are probably about 1000 votes a months

written by joedirt

| 10 months 2 weeks ago | CH

▲ 0 ▼ | flag spam (0)

^ Yes, I'm sure the cause of the decline after Nov 4th was because of comment downvotes, not the end of the US Presidential election.

written by NetRunner

| 10 months 2 weeks ago | CH

▲ 5 ▼ | flag spam (0)

Hmmmm... Presidential election.... Holiday vacation..... You don't really need Scoob' and the Mystery Machine to figure this one out. C'mon joedirt, You run the election08 channel for chrissakes!

written by dystopianfuturetoday

| 10 months 2 weeks ago | CH

▲ 3 ▼ | flag spam (0)

Yeah, it's definitely the election that was a high point followed by the quick holiday drop-off. The lack of activity on the site was very evident during the holidays.

I'm rather surprised by the huge jump in comment downvotes, even doubling from Sept to Oct. It's a little discouraging. I guess it was just a matter of time that the original sensitivity to the comment downvote wore off.

Nice to see our registrations are pretty consistent month to month.

written by lucky760

| 10 months 2 weeks ago | CH

▲ 0 ▼ | flag spam (0)

I want to see the % videos that are made of cats and kittens graph

written by Farhad2000

| 10 months 2 weeks ago | CH

▲ 4 ▼ | flag spam (0)

Lucky, you should re-run this with just grabbing the top15 users and see if they make up 1/2 of the votes/comments.

I'm guessing there is the 90/10 rule going on.

written by joedirt

| 10 months 2 weeks ago | CH

▲ 0 ▼ | flag spam (0)

that would be interesting.

Also, in the interests of 'proper' stats, wonder if it wouldn't be sensible to 'normalize' the data by eliminating the sock puppets? Ah meh, we can do it in our heads.