Interesting Tumblr Stats

I think rootjazz's bot supports note scraping, so that might be easier than messing with Ruby and push/pull from Git if you don't know what you're doing.
 
1. Go back to the GitHub page and make sure you've installed the correct version of Ruby. I first scraped these accounts a few months back so I don't recall but I do remember the version was important.
2. The script requires about a dozen or so Ruby Gems which need to be installed individually. Annoying I know but it is what it is. You see in the error "could not find curb-0.8.5'? That's the Ruby Gem that needs to be installed. Also ensure it is the exact version that is being asked for. Then re-run the script and it will tell you the next Gem that you need to install.
Keep doing that until it works.

Must be the Ruby version. I have installed the Curb gem.

I think rootjazz's bot supports note scraping, so that might be easier than messing with Ruby and push/pull from Git if you don't know what you're doing.

Since this is free.... and a little Ruby can't seem to hurt.
 
Must be the Ruby version. I have installed the Curb gem.



Since this is free.... and a little Ruby can't seem to hurt.

Free and learning something, always a good mix!




The next step for me, will be to create a 'real time' scraper using PHP and the API to scrape users that have reblogged in the last couple of minutes so that they might be actually still be online when I follow them and they'll see the follow notification straight away. But that's a little while away as I working on Twitter scripts at the moment, but I will get back to it.
Does the API give you an "is online" value? Have never looked at the API and always gone with website automation myself, but your thread has got me thinking about adding more filtering options to the program and perhaps I should look into the API..


I understand that a lot of this is outside the skill level of a lot of people on here. I just thought that I might share my findings. Basically, the newer/cleaner your list is, the better followback that you'll get, the faster you will grow...

Main Useful Metrics
* Last active on tumblr (whether a post / a note etc - something to show they are using the site)
* How active are they - number of posts / reblogs / likes. Are they likely to interact with content, thus syndicating it to their streams

Brainstorming Useful metrics
* General niche of the blog - are they worth targeting - could pull this by checking most popular tags of last 20 posts


Anyone else got any input on how better target users based on filters / other metrics?
 
Free and learning something, always a good mix!





Does the API give you an "is online" value? Have never looked at the API and always gone with website automation myself, but your thread has got me thinking about adding more filtering options to the program and perhaps I should look into the API..




Main Useful Metrics
* Last active on tumblr (whether a post / a note etc - something to show they are using the site)
* How active are they - number of posts / reblogs / likes. Are they likely to interact with content, thus syndicating it to their streams

Brainstorming Useful metrics
* General niche of the blog - are they worth targeting - could pull this by checking most popular tags of last 20 posts


Anyone else got any input on how better target users based on filters / other metrics?

No, the API doesn't tell you who is currently online, Google "Tumblr API" it is very well documented and easy to understand if you have basic scripting knowledge. What you can do is choose a handful of big blogs in your niche and the API can return the most recent 50 notes (it tells you if a note is a reblog or a like) for each post. It also tells you exactly to the second when that note was made, so you can scrape for reblogs from a blog in your niche from the last x minutes, for example. Once you have your users you can hit the API again and it will tell you, how many posts and a bunch of other info about the user so you can decide if they are worth following or not.

BTW. A quick note on the API. I just finished writing some scripts for Twitter to follow/unfollow and post, all using the API and its working well, so I'll prob recode my Tumblr scripts for their API. The good thing about the API is you can scrape for users in real time prior to following people, and, you don't have to worry about all this "pretending to be human" rubbish that the other bots do, because clearly if you're using the API you're not a human. Not to mention the API is more reliable and uses around 5% of the bandwidth that the commercial bots do. And with Tumblr you dont need to phone verify your account to use the API.
 
Seogootwit How many unique users you get with your 75 profiles?

As of 5 minutes ago I have 90,027 followers across the 75 accounts, having grown 2,648 followers in the last 24 hours. All account have been running on average 46 days. Not sure if these numbers are good or not, perhaps someone can offer some comparative figures from their own experiences. It's always good to know if you are on track or not.
 
Seogootwit How many unique users you get with your 75 profiles?

Not sure what happened with the last post. Oops.

As of 5 minutes ago I have 90,027 followers across the 75 accounts, having grown 2,648 followers in the last 24 hours. All account have been running on average 46 days. Not sure if these numbers are good or not, perhaps someone can offer some comparative figures from their own experiences. It's always good to know if you are on track or not.
 
Not sure what happened with the last post. Oops.

As of 5 minutes ago I have 90,027 followers across the 75 accounts, having grown 2,648 followers in the last 24 hours. All account have been running on average 46 days. Not sure if these numbers are good or not, perhaps someone can offer some comparative figures from their own experiences. It's always good to know if you are on track or not.

I think you have about 950 active users that actually visit your blogs. It's good result. You should put google analytics code to track your visitors.
 
Very interesting data, thanks for sharing.

Do you use Tumblr exclusively? Or do you also use Wordpress?
 
Very useful information, thanks for investigating.

Have you got any of your 75 blogs suspended? Tumblr seems to be very arbitrary in this regard.
 
I tried tumblr 3months ago and kinda failed ,never gonna get back to it
 
Very useful information, thanks for investigating.

Have you got any of your 75 blogs suspended? Tumblr seems to be very arbitrary in this regard.

None. Been running, on average, 47 days per blog. Am very aggressive too. Following close to the limit every day and posting around 40-50 bits of content a day on each blog. Haven't monetized yet. Maybe that's why they're safe.
 
just write a spider that constantly searches tags in your niche and gets usernames from the notes. then add to db. when your follower bot grabs from the db make it grab the newest usernames. ex. order by 'index' desc
 
Hey , thanks man. Its great to have great people like you to share research data like this. Keep up the good work.
 
Cheers. I've got a lot out this forum so I'm happy to give a little back.
 
Back
Top