I cam across one third party API which provides organized twitter data as per user request.
example
http://api.frrole.com/v1/curated-content?location=India&contenttype=link&category=sports&orderby=popularity&minrt=10&resultcount=100&apikey=4534sdg34343
It give top 100 tweets from india, which contains link in tweets and falling into sports category and having minimum rt count =10.
Can someone tell me how does this work? It gives json data as result. Does this result comes directly via twitter streaming APIs or third party's own data base?
Is there any way for streaming API to do this?
This is likely coming from their database. They are consuming the data from the Streaming API or another API and then applying their categorization to it, then giving you the ability to access it from an API.
Some of the criteria above you could get from the Streaming API such as the filter_level (to find top Tweets) but you wouldn't be able to get that whole set of requirements (especially the sports categorization).
Related
I've been using the YouTube Analytics API (I'm using the Java library), to retrieve analytics data for a YouTube Channel - I've been filtering the API calls to specific videoIDs.
However the data for the same date range, when I look in YouTube Analytics (the web interface) seems to differ.
Data from the YouTube Web Interface:
Data from the API:
Does anyone have an explanation for this ?
It says from this SO post that it is an intended behavior of both the API and the Youtube Analytics web interface.
To support this similar post, based from the documentation - How video views are counted:
If you're looking at a video you uploaded, you can monitor your views
more closely using YouTube Analytics. However, keep in mind that the
Realtime report only shows estimates of potential view activity and
might not match the number you see on the watch page.
Also, there are called Frozen View Count where on some videos, the view count might seem frozen or not show all the views that you expect. Video views are algorithmically validated to maintain fair and positive experiences for content creators, advertisers, and users. To verify that views are real and accurate, YouTube may temporarily slow down, freeze, or adjust the view count, as well as discard low-quality playbacks.
I'm working on a application which would gather YouTube user's video data and create some meaningful data and metrics to help the creators market their videos better and expand their audience.
The problem is that since December 18, if I'm not wrong, this kind of practice is forbidden.
Can someone from Google comment and explain this change? Why can't I create metrics based on YouTube data, even if I visibly communicate that this is not data from YouTube?
For example: I would like to fetch users video description and tell what's the keyword density, how well is it prepared for SEO (in % or something).
And I guess that this new term destroys many businesses which are doing exactly that thing, creating meaningful data based on YouTube API. (Tubular, TubeBuddy, VidIQ).
Please! Anyone?
Assuming that I'm quite new to the YouTube Content ID API (and in general to YouTube Content ID), i'm looking for a method for get quickly the views and the earnings of a single asset.
Analyzing the YouTube Analytics API, I have not found anything making reference to the assets,
then I tried to base myself on the claims to obtain the data which I need. It's work, but, having to walk multiple pages, this takes many requests to the YouTube API server and it responds really slowly.
I would like, in practice, achieve a similar result:
I'm using PHP for this, but I do mostly GET requests directly basing on the documentation of YouTube Content ID API
I've found the solution:
Using the Youtube Analytics API:
ids=contentOwner==MY_CONTENT_OWNER_ID
start-date=my_start_date
end-date=my_end_date
metrics=estimatedMinutesWatched,averageViewDuration,averageViewPercentage, views,subscribersGained
dimensions=video
filters=claimedStatus==claimed
max-results=10
sort=-views
I can obtain the top 10 videos claimed sorted by views.
With the video ids, I can get the views, the earnings and the asset id using the ClaimSearch reference in YouTube Content ID API I can find the rest of the informations that I need.
EDIT:
There is a dimension missed on the official documentation: asset.
I've updated the query of Analytics API:
ids=contentOwner==MY_CONTENT_OWNER_ID
start-date=my_start_date
end-date=my_end_date
metrics=estimatedMinutesWatched,averageViewDuration,averageViewPercentage, views,subscribersGained
dimensions=asset
filters=claimedStatus==claimed
max-results=10
sort=-views
and it show directly the assets.
Is it possible to access historic Twitter data?
I want to write a function that accesses tweets which match certain keywords, like the FilterStream in the Twitter streaming API, but can access tweets which have been created historically.
No it is not.
All you can do, is use search method to get recent matching tweets. (little amount if keywords are unpopular)
Or connect to streaming API and wait (in little while you will have historic tweets)
is there any service from where we can download tweets?
UPDATE!!!
Googling for sometime gave me this result
a.) http://snap.stanford.edu/data/twitter7.html
b.) http://140kit.com/datasets
Yes, there is. It's called the Twitter API.
As we have access to limited tweets by Twitter-API, we should make use of third-party resellers like Topsy for just the past data, GNIP just for streaming data, or DataSift for both streaming data as well as past data.
You might also want to check the following sites:
http://www.infochimps.com/collections/twitter-census
http://www.tweetarchivist.com/
Twitter API allows provides partial results, it gives you the last 100 or even 500 tweets fo every search. If you need to keep tweets long term, twitter API shows its limits.
I had same need as you apparently hae and I developed a tool that queries twitter API periodically and stores search results on a Wordpress database.
I called the tools twittcorder and you can find a live demo on twittcorder.com
I hope this helps.
These other data sources are probably shared against the Twitter TOS. I wouldn't want to invest my time and effort building something on datasets that are non-repeatable. The Twitter Streaming API allows collection of a sample of Tweets.
There's also Gnip: http://gnip.com/.
Sysomos is there for complete data analysis including twitter, faecbook and various boards and forums