[STATS] Bitcointalk stats vs BTC value chart - UPDATE 01/09 - analysis

[dnaleor]

"Smooth chart" ( average of day-1, current_day and day+1)

Logarithmic version:


Based on the stats of this forum, I created a nice chart.
What we can see, it correlates nicely with the bitcoin price.

However, some other conclusions can be made:
1) the number of new posts stays relatively high compared to the previous bitcoin price spike in june 2011.
The data is still "fresh" but if this trend continues, we could say that the overall interest ('talk') about bitcoin has gone up and will stay at this high level.
2) a lot of new members during price rally's, but this trend never continues. So the number of new members could maybe indicate a bubble effect?
3) the number of page views hit new records after the price spike. Sadly, we can't compare to the june 2011 price spike (no data). What could be the explanation?

Any thoughts?


(BTC price added. Used data weighted average MTGOXusd from bitcoincharts.com)

[ashaw596]

I believe a LARGE part of the booms come because its now summer.

[kodo]

We are here to stay, definitly.

[bluemeanie1]

can you plot the BTC/USD exchange rate on top of that graph?

-bm

[raze]

Just remember correlation doesn't imply causation.

[dnaleor]

can you plot the BTC/USD exchange rate on top of that graph?

-bm

yes, if someone has an excel sheet with the daily BTC-price

[N12]

I believe a LARGE part of the booms come because its now summer.

Sell in May and go away.

[bluemeanie1]

can you plot the BTC/USD exchange rate on top of that graph?

-bm

yes, if someone has an excel sheet with the daily BTC-price

try this: https://docs.google.com/spreadsheet/ccc?key=0AmcTCtjBoRWUdHJuUE1mUkFxa3A0eHBDQkxZLVVFZmc#gid=0

[zby]

Based on the stats of this forum, I created a nice chart.
What we can see, it correlates nicely with the bitcoin price.

However, some other conclusions can be made:
1) the number of new posts stays relatively high compared to the previous bitcoin price spike in june 2011.
The data is still "fresh" but if this trend continues, we could say that the overall interest ('talk') about bitcoin has gone up and will stay at this high level.
2) a lot of new members during price rally's, but this trend never continues. So the number of new members could maybe indicate a bubble effect?
3) the number of page vieuws hit new records after the price spike. Sadly, we can't compare to the june 2011 price spike (no data). What could be the explanation?

Any thoughts?


(If you want to do data mining, check the excel-file here: http://www.sendspace.com/file/atkufw
And maybe donate something for the effort )

bots

bitcointalk is now a target

[dnaleor]

can you plot the BTC/USD exchange rate on top of that graph?

-bm

yes, if someone has an excel sheet with the daily BTC-price

try this: https://docs.google.com/spreadsheet/ccc?key=0AmcTCtjBoRWUdHJuUE1mUkFxa3A0eHBDQkxZLVVFZmc#gid=0

Ill update the chart tonight, Thx :-)

[dnaleor]

updated the firstpost.

I also have a "smooth" version (I used for each date the average from the previous, the current and the next day; this is done except for the BTC value)




do you like this one better? (better readability?)

[bluemeanie1]

that's really impressive.

the owners of this forum could use this data to hedge the price of BTC.

[puffpuffpass]

the new threads, new members, new posts, etc. seem to come after the fact, proving that this forum is a joke.

[kayrice]

I love the ignore button. Instead of replying to people I get to look next to their name, see bright orange, and know I don't have to waste my time explaining something to the person above me.

[funktion]

thanks, this is pretty interesting. the correlation does follow nicely at the beginning but totally collapses after the big bubble burst. now the trick is to guess whether increased activity means drop or rise. it would be cool if an analysis of posts was done looking for positive/negative word indicators and see how that correlates with the price.

[hello_good_sir]

thanks, this is pretty interesting. the correlation does follow nicely at the beginning but totally collapses after the big bubble burst. now the trick is to guess whether increased activity means drop or rise. it would be cool if an analysis of posts was done looking for positive/negative word indicators and see how that correlates with the price.

I'm pretty sure that the chart does not show a decrease in correlation.  This chart is yet another example of why all bitcoin price charts need to be log-scaled.  When people look at linear-scaled charts they draw incorrect conclusions.

[Luckybit]

Based on the stats of this forum, I created a nice chart.
What we can see, it correlates nicely with the bitcoin price.

However, some other conclusions can be made:
1) the number of new posts stays relatively high compared to the previous bitcoin price spike in june 2011.
The data is still "fresh" but if this trend continues, we could say that the overall interest ('talk') about bitcoin has gone up and will stay at this high level.
2) a lot of new members during price rally's, but this trend never continues. So the number of new members could maybe indicate a bubble effect?
3) the number of page vieuws hit new records after the price spike. Sadly, we can't compare to the june 2011 price spike (no data). What could be the explanation?

Any thoughts?


(If you want to do data mining, check the excel-file here: http://www.sendspace.com/file/atkufw
And maybe donate something for the effort )

bots

bitcointalk is now a target

I think you're right. Bitcointalk was DDOSed as part of the same effort to shut down the exchanges.

[Luckybit]

DDOS from bots would be the cause of both the crash and the spike on Bitcointalk.

[dnaleor]

DDOS from bots would be the cause of both the crash and the spike on Bitcointalk.

did the bots create the accounts and posted those messages? Don't think so.
Maybe the pageviews can be explained by google bots and/or ddos

[dnaleor]

updated the charts:

"Smooth chart" ( average of day-1, current_day and day+1): http://img515.imageshack.us/img515/3995/vg9w.png

Logarithmic version:

"Smooth" logaritmic version: http://img826.imageshack.us/img826/6946/wxa.png

Any comments? The logaritmic verion is really interesting. You can see that the current "bubble" maybe was triggered by the number of posts and members increasing.
The previous bubble triggered a higher number of posts and members. So it's the other way around.

This chart makes me even more bullish on BTC!